relationships between nitrogen transformation rates and gene abundance in a riparian buffer soil

14
Relationships Between Nitrogen Transformation Rates and Gene Abundance in a Riparian Buffer Soil Lin Wu Deanna L. Osmond Alexandria K. Graves Michael R. Burchell Owen W. Duckworth Received: 7 February 2012 / Accepted: 20 July 2012 / Published online: 22 August 2012 Ó Springer Science+Business Media, LLC 2012 Abstract Denitrification is a critical biogeochemical process that results in the conversion of nitrate to volatile products, and thus is a major route of nitrogen loss from terrestrial environments. Riparian buffers are an important management tool that is widely utilized to protect water from non-point source pollution. However, riparian buffers vary in their nitrate removal effectiveness, and thus there is a need for mechanistic studies to explore nitrate dynamics in buffer soils. The objectives of this study were to examine the influence of specific types of soluble organic matter on nitrate loss and nitrous oxide production rates, and to elucidate the relationships between these rates and the abundances of functional genes in a riparian buffer soil. Continuous-flow soil column experiments were performed to investigate the effect of three types of soluble organic matter (citric acid, alginic acid, and Suwannee River dis- solved organic carbon) on rates of nitrate loss and nitrous oxide production. We found that nitrate loss rates increased as citric acid concentrations increased; however, rates of nitrate loss were weakly affected or not affected by the addition of the other types of organic matter. In all experiments, rates of nitrous oxide production mirrored nitrate loss rates. In addition, quantitative polymerase chain reaction (qPCR) was utilized to quantify the number of genes known to encode enzymes that catalyze nitrite reduction (i.e., nirS and nirK) in soil that was collected at the conclusion of column experiments. Nitrate loss and nitrous oxide production rates trended with copy numbers of both nir and 16s rDNA genes. The results suggest that low-molecular mass organic species are more effective at promoting nitrogen transformations than large biopolymers or humic substances, and also help to link genetic potential to chemical reactivity. Keywords Riparian buffer Nitrogen Denitrification Quantitative PCR Nitrous oxide Nitrate Introduction Riparian buffers are an important management tool for reducing the impact of agricultural nonpoint source nitrate (NO 3 - ) pollution on water quality. These buffer strips of uncultivated land bordering drainages have proven effec- tive at reducing NO 3 - fluxes to surface waters; however, buffers vary widely in their NO 3 - removal efficacy (Hill 1996; Smith and others 2006; Lowrance and others 1997; Osborne and Kovacic 1993; Mayer and others 2007). At the Center for Environmental Farming (CEFS; Goldsboro, NC), which contains a long term monitoring site associated with this study (Dukes and others 2002; King 2006; Knies 2009), the type of vegetation grown on riparian buffers has been implicated as a key factor in controlling buffer effectiveness, although the mechanism of action has not been conclusively demonstrated (King 2006). Riparian vegetation also has been widely reported to be a key factor in controlling pollutant dynamics, partially by providing L. Wu D. L. Osmond A. K. Graves O. W. Duckworth (&) Department of Soil Science, North Carolina State University, Raleigh, NC 27695-7619, USA e-mail: [email protected] Present Address: L. Wu Department of Statistics, University of North Carolina, Chapel Hill, NC, USA M. R. Burchell Department of Biological and Agricultural Engineering, North Carolina State University, Raleigh, NC 27695-7625, USA 123 Environmental Management (2012) 50:861–874 DOI 10.1007/s00267-012-9929-z

Upload: alexandria-k-graves

Post on 13-Dec-2016

214 views

Category:

Documents


0 download

TRANSCRIPT

Relationships Between Nitrogen Transformation Rates and GeneAbundance in a Riparian Buffer Soil

Lin Wu • Deanna L. Osmond • Alexandria K. Graves •

Michael R. Burchell • Owen W. Duckworth

Received: 7 February 2012 / Accepted: 20 July 2012 / Published online: 22 August 2012

� Springer Science+Business Media, LLC 2012

Abstract Denitrification is a critical biogeochemical

process that results in the conversion of nitrate to volatile

products, and thus is a major route of nitrogen loss from

terrestrial environments. Riparian buffers are an important

management tool that is widely utilized to protect water

from non-point source pollution. However, riparian buffers

vary in their nitrate removal effectiveness, and thus there is

a need for mechanistic studies to explore nitrate dynamics

in buffer soils. The objectives of this study were to

examine the influence of specific types of soluble organic

matter on nitrate loss and nitrous oxide production rates,

and to elucidate the relationships between these rates and

the abundances of functional genes in a riparian buffer soil.

Continuous-flow soil column experiments were performed

to investigate the effect of three types of soluble organic

matter (citric acid, alginic acid, and Suwannee River dis-

solved organic carbon) on rates of nitrate loss and nitrous

oxide production. We found that nitrate loss rates increased

as citric acid concentrations increased; however, rates of

nitrate loss were weakly affected or not affected by the

addition of the other types of organic matter. In all

experiments, rates of nitrous oxide production mirrored

nitrate loss rates. In addition, quantitative polymerase chain

reaction (qPCR) was utilized to quantify the number of

genes known to encode enzymes that catalyze nitrite

reduction (i.e., nirS and nirK) in soil that was collected at

the conclusion of column experiments. Nitrate loss and

nitrous oxide production rates trended with copy numbers

of both nir and 16s rDNA genes. The results suggest that

low-molecular mass organic species are more effective at

promoting nitrogen transformations than large biopolymers

or humic substances, and also help to link genetic potential

to chemical reactivity.

Keywords Riparian buffer � Nitrogen � Denitrification �Quantitative PCR � Nitrous oxide � Nitrate

Introduction

Riparian buffers are an important management tool for

reducing the impact of agricultural nonpoint source nitrate

(NO3-) pollution on water quality. These buffer strips of

uncultivated land bordering drainages have proven effec-

tive at reducing NO3- fluxes to surface waters; however,

buffers vary widely in their NO3- removal efficacy (Hill

1996; Smith and others 2006; Lowrance and others 1997;

Osborne and Kovacic 1993; Mayer and others 2007). At

the Center for Environmental Farming (CEFS; Goldsboro,

NC), which contains a long term monitoring site associated

with this study (Dukes and others 2002; King 2006; Knies

2009), the type of vegetation grown on riparian buffers has

been implicated as a key factor in controlling buffer

effectiveness, although the mechanism of action has not

been conclusively demonstrated (King 2006). Riparian

vegetation also has been widely reported to be a key factor

in controlling pollutant dynamics, partially by providing

L. Wu � D. L. Osmond � A. K. Graves � O. W. Duckworth (&)

Department of Soil Science, North Carolina State University,

Raleigh, NC 27695-7619, USA

e-mail: [email protected]

Present Address:L. Wu

Department of Statistics, University of North Carolina,

Chapel Hill, NC, USA

M. R. Burchell

Department of Biological and Agricultural Engineering, North

Carolina State University, Raleigh, NC 27695-7625, USA

123

Environmental Management (2012) 50:861–874

DOI 10.1007/s00267-012-9929-z

different types of organic carbon (C) that may affect soil

biogeochemistry in the riparian zone (Dosskey and others

2010).

The riparian zone has been identified as an area that may

have particularly high rates of denitrification (Hedin and

others 1998). Although denitrification is notoriously diffi-

cult to assess because of the nature of the products and the

complicated web of environmental variables that partially

control the process (Firestone and Davidson 1989; Knowles

1982; Myrold and Tiedje 1985; Nommik 1956), it is gen-

erally accepted that availability of NO3- and organic C are

the two major limiting factors for denitrification under

anaerobic conditions (Greenan and others 2006; Rivett and

others 2008; Starr and Gillham 1993; Hill and others 2000;

Bradley and others 1992). Significant relationships have

been observed between denitrification rates and various

operationally defined C fractions, including soluble,

extractable, and mineralizable C (Beauchamp and others

1980; Bijay-Singh and Whitehead 1988; Burford and

Bremner 1975; Davidson and others 1987; Hill and Cardaci

2004; Myrold and Tiedje 1985; Stanford and others 1975;

McCarty and Bremner 1992; Pintar and Lobnik 2005;

Hernandez and Mitsch 2007). However, which fractions of

soil organic matter correlate with reactivity differs between

studies, emphasizing the variability in the bioavailability of

soluble C and the possible influence of other edaphic and

environmental factors on the process. Because many con-

voluting factors control denitrification, continuous-flow soil

column experiments, which have been successfully used to

quantify rates of NO3- loss and nitrous oxide (N2O) pro-

duction at specifically defined chemical and hydrological

conditions, are a powerful tool to study the effects of C

concentration and type on denitrification (Knies 2009;

Pavel and others 1996; Willems and others 1997). Although

it is difficult to truly decouple N transformations, these

columns allow for the isolation of environmental variables

and control of the type of organic C present.

Molecular biology techniques provide insights into

denitrification that may be complementary to chemical

measurements of N transformations (Wallenstein and others

2006). Biological denitrification involves four enzyme

catalyzed reductions: NO3- ? NO2

- ? NO ? N2O ?N2. A key step in the denitrification process is the con-

version of NO2- to NO (nitric oxide), which is the first step

in the reaction sequence that produces a gaseous product

that can be volatized and exported from the buffer system.

This reaction is catalyzed by enzymes encoded by nitrite

(NO2-) reductase genes (i.e., nirK and nirS) that can be

exploited as molecular biomarkers to study biological

denitrification (Wallenstein and others 2006; Throback and

others 2004; Philippot 2002). Although studies have sug-

gested that denitrification rates trend with bacterial counts

(Miller and others 2012; Ardakani and others 1975;

Bowman and Focht 1974; Myrold and Tiedje 1985), others

have not seen a clear relationship between denitrification

rate and microbial community composition (Henry and

others 2008; Mounier and others 2004; Groffman and others

2006). Thus, the relationship between chemical denitrifi-

cation rate and microbial activity is poorly understood.

In the present study, column experiments were utilized

to evaluate NO3- loss and N2O production rates in a

riparian buffer soil as a function of dissolved organic C

type and concentration. In addition, we utilized quantitative

PCR to measure the copy number of nirS and nirK func-

tional genes to explore the relationship between N trans-

formation rates and gene abundance. The specific

objectives of this study were to: (1) determine the effects of

specific organic molecules on rates of N reduction (NO3-

loss and N2O oxide production rates) and (2) quantitatively

link this chemical reactivity to genetic potential (gene copy

numbers) in a riparian buffer soil. The results provide

specific information about the effects of different classes of

organic molecules on the biology and chemistry of N

transformations in anaerobic soils.

Materials and Methods

Soil Materials

A soil was collected in May 2009 from a fescue-vegetated

riparian buffer located adjacent to a drainage ditch at the

Center for Environmental Farming Systems (CEFS) in

Goldsboro, North Carolina. The buffer vegetation was

established in 1998, with the adjacent field historically used

as a cattle pasture and spray field for swine waste (Dukes

and others 2002). This site has been the subject of a long-

term nutrient dynamics monitoring project (Dukes and

others 2002; King 2006; Knies 2009). Soil was collected

from the riparian buffer by using a hand auger from a depth

of 3 m, which is the location of the aquifer that was the

subject of monitoring studies. Soil samples were kept on

ice during transportation and stored at 4 �C in the dark until

packed into soil columns. Based on the county soil map,

buffer soil was classified as Lumbee sandy loam (Barnhill

and others 1974). Our analysis of texture via the hydrom-

eter method and sieving (Gee and Or 2002) revealed the

soil to be a sand (96.7 % sand, 1.0 % silt, 2.3 % clay).

Total N and organic C concentrations were 0.02 % and

0.07 %, respectively, as determined by Perkin Elmer 2400

CHNS elemental analyzer.

Experimental Design

All solutions used in the study were made with type I

deionized water (DI) with a resistivity of 18.3 MX cm. All

862 Environmental Management (2012) 50:861–874

123

influent solutions were boiled and then continuously

purged with humidified argon (Ar) gas to exclude oxygen.

A NO3- concentration of 5.0 mg N L-1 (as KNO3) was

used in all experiments. This value was chosen because it is

an environmentally relevant concentration that is repre-

sentative of NO3- concentrations typically found in aqui-

fers at CEFS (King 2006; Knies 2009). Field measurements

have suggested that organic C availability limits denitrifi-

cation in these buffers (Knies 2009). Thus, influent solu-

tions also contained different concentrations of three

different organic C sources, as described below.

Three types of organic matter were utilized as model

compounds to represent broader families of organic mol-

ecules. Citric acid (C6H8O7, Acros Organics), a low

molecular mass organic acid (LMMOA), was chosen to

represent metabolites that are commonly exuded by plants

(Jones 1998; Curl and Truelove 1986; Jones and others

2003). Alginic acid (approximately [C6H8O6]n, Acros

Organics), a large macromolecule composed of uronic

sugar acids, was selected to represent environmentally

common microbial polysaccharides (Perry and others 2006;

Perry and others 2004). To represent soluble humic sub-

stances, Suwannee River dissolved organic carbon

(SRDOC) was obtained from International Humic Acid

Society (Suwannee River natural organic matter sample

1R101 N) and used without further purification. This

reference material has been extensively characterized,

with details available on the supplying society’s website

(http://www.ihss.gatech.edu/). Most notably, the SRDOC

has been found to be composed of 52.47 % C and 1.10 %

N by mass. For each type of organic molecule, experiments

were conducted with 4.0 mg C L-1, 8.0 mg C L-1, and

16.0 mg C L-1. A control experiment was also conducted

with no C added to the solution. Initial solution pH was 5.7

and 5.5 for alginic acid and SRDOC. To keep pH constant

between trials and avoid confounding effects from low pH

(Groffman and others 1991), inlet solutions containing

citric acid were adjusted to pH 6.0 by addition of NaOH.

Sampling and Analysis

Columns were constructed from polyvinyl chloride (PVC)

pipe that was 13.2 cm long and 3.0 cm in diameter. Soil

was packed into columns to a bulk density of 1.6 g cm-3, a

representative value for soil at CEFS (Dukes 2000). Col-

umns were sealed with rubber stoppers lined with fiberglass

fabric and equipped with plastic tubing outlets. Stoppers

were attached with waterproof polyurethane-based glue to

columns to ensure a tight fit.

Anaerobic continuous flow column experiments were

conducted to measure the rates of NO3- loss and N2O

production in the presence of different forms and concen-

trations of organic C (Pavel and others 1996; Knies 2009).

Boiled and Ar-purged solutions were pumped through soil

columns via fluorinated ethylene propylene (FEP) tubing

wrapped in aluminum foil to exclude atmospheric gases,

except for a short stretch of tygon tubing that was within a

peristaltic pump (Manostat). These precautions minimized

contamination from atmospheric gases, thus keeping col-

umns anaerobic. All experiments were conducted at

ambient laboratory temperature (25 ± 2 �C).

For each C type and concentration, columns experi-

ments were conducted for six days, which was sufficient

for the effluent to reach a steady-state concentration of

dissolved NO3- (cf. days 3-6 in Fig. 1a); all experiments

were conducted in triplicate. Each day, approximately

30 ml of effluent was collected in 50 ml disposable plastic

beakers over a timed interval (ca. 30 min). The mass of

effluent was measured to determine specific flow rate,

0

1

2

3

4

5

6

[NO

3- ] (m

g N

L-1)

Time (day)

0

1

2

3

4

5

6

Time (day)

0

1

2

3

4

5

6

1 2 3 4 5 6

1 2 3 4 5 6

1 2 3 4 5 6

Time (day)

a

b

c

[NO

3- ] (m

g N

L-1)

[NO

3- ] (m

g N

L-1)

influent

effluent of triplicate columns

influent

effluent of triplicate columns

influent

effluent of triplicate columns

Fig. 1 NO3- concentrations in the influent (9) and effluent (open

markers) in experiments containing 8 mg C L-1 a citric acid, b alginic

acid, and c SRDOC. Lines (solid = influent and dashed = effluent)

are presented to guide the eye and do not represent a model fit

Environmental Management (2012) 50:861–874 863

123

which was 19.0–32.3 ml h-1 for all experiments. The flow

rate was chosen to optimize residence time such that the

system would have a reasonable limit of detection for

NO3- loss rate, but would keep the steady-state NO3

-

concentration from becoming depleted ([NO3-] [

2 mg N L-1). The pH of effluent samples was measured

using an Accumet Excel pH/conductivity meter (XL20).

During experiments containing alginic acid and SRDOC,

pH change from the influent to effluent was pH \ 0.5; for

citric acid experiments, the change in pH = 0.3–1.0. In all

cases, pH drift was upward. Effluent samples were filtered

by syringe through 0.22 lm nylon filters (Millipore), fro-

zen, and stored at -20 �C for further analysis.

Effluent from each column (ca. 5 ml) was collected via a

hypodermic needle into a sealed vacutainer tube (Becton,

Dickinson and Company; 22.4 ml capacity) for determi-

nation of N2O. Attempts to utilize evacuated vacutainers

were unsuccessful because the vacuum produced suction

on the column outlet that increased flow rates and fre-

quently caused column clogging, presumably due to the

dislodging of soil particles. Therefore, solution was col-

lected into air-filled vacutainers and the background con-

tribution from air was subtracted, as described below.

Samples were analyzed for NO3-, NO2

-, and N2O to

determine rates of NO3- loss and N2O production. For

determination of NO3- loss, the NO3

- concentrations of

the influent and effluent solutions were determined color-

metrically with a sulfanilamide colorimetric reagent using

flow injection analysis (QuikChem Method 10107-01-A,

Lachat Instruments QuikChem 8000) with detection limit

of 0.05 mg N L-1. In experiments with measureable NO3-

loss, NO2- concentration in the effluent was also measured

colormetrically via the same method as NO3-. The NO2

-

concentrations were less than 1.00 mg N L-1 in all cases,

with most below the detection limit of 0.05 mg N L-1

(data not shown).

For analysis of N2O production, 10 ml of gas was man-

ually extracted from the headspace of vacutainer tube with a

gastight syringe and injected into a gas chromatograph

(Hewlett Packard 5890 GC-ECD, injector tempera-

ture = 60 �C). A Henry’s Law calculation (KH = 0.0257

M atm-1; Stumm and Morgan 1996) indicated that ca. 85 %

of N2O will be present in the headspace gas phase at equi-

librium. It should be noted that this calculation was specific

to volume of both the container and water sample, and cannot

necessarily be generalized to others systems. The concen-

tration of dissolved N2O resulting from partitioning from the

effluent in the gas phase was determined by difference from

the ambient background concentration (320 ppbV), thus

yielding an effective detection limit of 50 ppbV. The con-

centration of N2O that was dissolved in the effluent was then

calculated by assuming the headspace gas concentration was

originally dissolved in the solution. We acknowledge that

this approach contains an approximation (i.e., that gas phase

N2O accounts for all N2O in the system) that may result in a

small systematic underreporting of N2O production.

The total dissolved C concentration in effluent samples

from citric acid treatments was measured with TOC/N

Analyzer (Shimadzu Scientific Instruments, Japan). The

average effluent dissolved C concentration was

2.6 mg L-1, 4.0 mg L-1, and 4.7 mg L-1 for experiments

initially containing 4.0 mg C L-1, 8.0 mg C L-1, and

16.0 mg C L-1 citric acid, respectively. These values

suggest that the C concentration was not completely

depleted in the column by microbial processes.

Calculation of NO3- Loss and N2O Production Rates

The overall mass-normalized NO3- loss rate (RNO�

3;

mg N h-1 g-1) was calculated using the following equa-

tion (Knies 2009; Pavel and others 1996):

RNO�3¼

q� D NO�3� �

mð1Þ

where q was flow rate (L h-1), D[NO3-] was the change of

NO3- concentration (mg N L-1) between inflow and

effluent, and m was the mass of soil packed into the column

(g). The rate calculation was based on the steady-state

concentration, which typically was the final 4-5 days of the

experiment. The limit of detection (LOD) of

RNO�3

= 4 9 10-5 mg N h-1 g-1 was calculated via Eq. 1

by determining the minimum reliable difference between

influent and effluent (0.1 mg N L-1).

Similarly, the N2O production rate (RN2O;

mg N h-1 g-1) was calculated as follows:

RN2O ¼q� D N2O½ �

mð2Þ

where q was the flow rate (L h-1), DC (N2O) was the

concentration of N2O dissolved in column effluent

(mg N L-1), and m was the initial mass of soil in the

column (g). The LOD for N2O production was estimated to

be RN2O = 10-8 mg N h-1 g-1 loss rate, as calculated via

Eq. 2 from the detection limit against background. For both

NO3- loss and N2O production rates, uncertainty was

reported as twice the standard deviation (±1 standard

deviation) of three replicates of experiments.

Quantitative PCR

At the conclusion of experiments, soil samples were col-

lected in beakers, homogenized by mechanical mixing with

a sterile spatula, and stored at 4 �C for no more than

2 weeks, at which point DNA was extracted from soil in

each column. The PowerSoilTM

DNA Isolation Kit (Mo BIO

Laboratories, USA) was used to extract and purify genomic

864 Environmental Management (2012) 50:861–874

123

DNA from 1.0 g homogenized soil (wet weight) from each

soil column; this procedure was completed twice for each

column. Extracted DNA was quantified by spectropho-

tometry at 260 nm using a Nanodrop spectrophotometer

(Thermo scientific, USA). DNA was stored at -20 �C until

amplified by quantitative PCR (qPCR).

Previously published primers were used in this study to

amplify 16 rDNA and nir genes (Kandeler and others 2006;

Throback and others 2004; Henry and others 2004). The

primer sequence and annealing temperature for corre-

sponding targets are shown in Table 1. Amplification of

qPCR products conducted by using a MasterCycler� EP

Realplex cycler (Eppendorf, USA).

Amplifications were performed in 25 lL reaction mix-

tures using QuantiFastTM

SYBR� Green PCR Kit (QIA-

GEN, Valencia, CA). The reaction mixture contained

2.5 lL of each primer for gene (1 lM final concentration

for each), 12.5 lL of 2 9 master mix (including HotStar

Taq Plus DNA polymerase, QuantiFast SYBR Green PCR

buffer, dNTP mix with dUTP, SYBR Green I, ROX dye

and 5 mM MgCl2), 2 lL of DNA template (ca. 10 ng total

DNA), and RNase-free water to complete the 25 lL

volume.

Thermocycler conditions for nirS qPCR were 300 s at

95 �C for enzyme activation as recommended by the

manufacturer (QuantiFastTM

SYBR� Green PCR Kit,

QIAGEN), followed by 35 cycles of 45 s at 94 �C for

denaturation, 45 s at 57 �C for annealing, and 60 s at

72 �C for extension and data acquisition. The reaction

was completed after 300 s at 72 �C for final extension and

a final temperature gradient step from 60 �C to 95 �C

with an increase of 0.2 deg s-1 used to obtain a specific

denaturation curve. The conditions for nirK and 16s

rDNA qPCR were similar as those for nirS, except that

the annealing temperatures were 58 �C and 53 �C,

respectively (Henry and others 2004; Nogales and others

2002). Quantitative PCR efficiencies for were 91 %,

101 %, and 94 % for 16s rDNA, nirK, and nirS genes,

respectively.

The purity of amplified products was verified by the

observation of a single melting curve or the presence of a

unique band of the expected size in a 2 % (weight base)

agarose gel stained with ethidium bromide. Standard

curves for absolute quantitation were obtained from serial

dilutions of a known amount of genomic DNA (obtained

from the American Type Culture Collection; ATCC) con-

taining a fragment of the 16s rDNA (ATCC8750, 3285 kb),

nirS (ATCC 17699D-5, 4052 kb), or nirK (ATCC8750,

3285 kb) gene. The critical threshold (Ct) values were

plotted as a log function of the target DNA copy numbers.

Tenfold serial dilutions of standard genomic DNA ranging

from 102 to 106 gene copies were used in triplicate as a

template to determine the standard curve. Three points of

the standards were used as positive controls in each reac-

tion. A reaction mixture with no template was used as a

negative control. Uncertainty was determined as the stan-

dard deviation of nine replicate amplifications for each

trial.

Statistical Analysis

All statistical calculations were performed in SAS 9.2

(Cary, NC). Two-way analyses of variance (ANOVA) were

performed to compare the effects of C type and concen-

tration on NO3- loss rates, N2O production rates, and nirK,

nirS, and 16S rDNA gene abundances. Means were com-

pared using the Fisher’s least significant difference (LSD)

test at significant level P \ 0.05. For NO3- loss and N2O

production rates and all gene abundances, the C concen-

tration, C type, and a two-way interaction were significant.

The two-way interactions are evident in the data, which

show that all rates and gene abundances trend positively

with C concentration but increase at differing rates (cf.

Tables 2, 3; Figs. 2, 3). The physical rationale for the

different responses to C type is explored in the Discussion

Section. Additionally, one-way analyses of variance

(ANOVA) were performed to compare the gene abun-

dances and N transformation rates both as a function of C

Table 1 Primer sets used for amplification of nir and 16s rDNA (Kandeler and others 2006; Throback and others 2004; Henry and others 2004)

Gene Amplicon size (bp) Primer Annealing

temperature (�C)Name Sequence (50-30)

nirS 425 nirSCd3aF AACGYSAAGGARACSGGa 57

nirSR3cd GASTTCGGRTGSGTCTTSAYGAA

nirK 165 nirK876 ATYGGCGGVAYGGCGA 60

nirK1040 GCCTCGATCAGRTTRTGGTT

16s 200 Eub338 ACTCCTACGGGAGGCAGCAG 53

Eub518 ATTACCGCGGCTGCTGG

a Y was a mixture of C and T. S was a mixture of C and G. R was a mixture of A and G

Environmental Management (2012) 50:861–874 865

123

concentration and type. Again, means were compared

using the Fisher’s least significant difference (LSD) test at

significant level P \ 0.05.

Results

Rates of Nitrate Loss

Examples of data from column experiments, with con-

centrations of NO3- in effluent from triplicate column

experiments plotted as a function of time, are shown in

Fig. 1. Large differences in the rates of NO3- loss were

observed in experiments that utilize different C types. In

Fig. 1A, decreases in NO3- concentration (up to

2.1 mg N L-1) were observed in citric acid treatments at

the concentration of 8 mg C L-1. However, alginic acid

(Fig. 1B) and SRDOC (Fig. 1C) showed no detectable

decrease of NO3- concentrations in the effluent.

Similar results were also evident in experiments at other

concentrations. The average NO3- loss rates (as calculated

from Eq. 1) for all combinations of C type and concen-

tration are shown in Fig. 2a and Table 2. Rates of NO3-

loss increased with increasing citric acid concentration

(R2 = 0.98). Alginic acid had a statistically significant

effect on NO3- loss rates in experiments containing 4 and

16 mg C L-1; however, all rates were significantly less

than those promoted by citric acid at corresponding con-

centrations. In control experiments with no C added, as

well as in all SRDOC treatments, only small changes in

Table 2 The NO3- loss rate and N2O production rate (mg N h-1 g-1) for given combination of carbon type and concentration

Rate Concentration

(mg C L-1)

Rate (mg N h-1 g-1)

Citric Acid Alginic Acid SRDOC

NO3- loss 0 5 ± 6 9 10-5A,a 5 ± 6 9 10-5A,a 5 ± 6 9 10-5A,a

4 1.0 ± 0.8 9 10-4A,b 4 ± 3 9 10-5B,b LOD

8 4 ± 1 9 10-4A,c LOD LOD

16 8 ± 2 9 10-4A,d 4 ± 2 9 10-5B,b LOD

N2O production 0 4 ± 1 9 10-7A,a 4 ± 1 9 10-7A,a 4 ± 1 9 10-7A,a

4 4 ± 3 9 10-7A,a 3.3 ± 0.9 9 10-7A,a 4 ± 2 9 10-7A,a

8 2 ± 1 9 10-6A,b 4 ± 1 9 10-7B,a 3 ± 1 9 10-7B,a

16 3.8 ± 0.9 9 10-6A,c 2.3 ± 0.9 9 10-7B,b 5 ± 2 9 10-7B,b

For NO3- loss rate, limit of detection (LOD)\4 9 10-5 mg N h-1 g-1, based on the reliable detection limit of NO3

- loss (0.1 mg N g-1). For

N2O production rate, LOD\10-8 mg N h-1 g-1, based on detection limit of N2O detection against background (50 ppbV). For both NO3- loss

and N2O production rates, uncertainty is reported as twice the standard deviation (±1 standard deviation) of three replicates of experiments. For

NO3- loss and N2O production rates, values followed by the same capital letter in a row or a lowercase letter within a column are not

significantly different (P [ 0.05) based on a LSD test

Table 3 Gene copy numbers for different carbon types as a function of concentration

Gene Concentration

(mg C L-1)

Gene copy numbers (copy g-1 soil)

Citric acid Alginic acid SRDOC

16s rDNA 0 6 ± 3 9 106A,a 6 ± 3 9 106A,a 6 ± 3 9 106A,a

4 8 ± 3 9 107A,b 2 ± 3 9 107B,b 1.4 ± 0.4 9 107B,b

8 1.2 ± 0.4 9 108A,b 2.7 ± 0.7 9 107B,b 1.4 ± 0.3 9 107B,b

16 1.8 ± 0.3 9 108A,c 3 ± 1 9 107B,b 1.3 ± 0.3 9 107B,b

nirK 0 5 ± 4 9 106A,a 5 ± 4 9 106A,a 5 ± 4 9 106A,a

4 7 ± 3 9 107A,b 1.7 ± 0.9 9 107B,a 1.2 ± 0.5 9 107B,a

8 7 ± 2 9 107A,b 2.3 ± 0.8 9 107B,a 1.2 ± 0.3 9 107B,a

16 9 ± 4 9 107A,b 2 ± 1 9 107B,a 1.0 ± 0.3 9 107B,a

nirS 0 2 ± 2 9 106A,a 2 ± 2 9 106A,a 2 ± 2 9 106B,a

4 3 ± 1 9 107A,b 7 ± 4 9 106B,b 6 ± 3 9 106B,a

8 4 ± 1 9 107A,b 1.3 ± 0.7 9 107B,b 9 ± 6 9 106B,a

16 4 ± 2 9 107A,b 1.2 ± 0.7 9 107B,b 6 ± 3 9 106B,a

Uncertainty is reported as twice the standard deviation (± 1 standard deviation) of nine amplifications at each condition. For a gene, values

followed by the same capital letter in a row or a lowercase letter within a column are not significantly different (P [ 0.05) based on a LSD test

866 Environmental Management (2012) 50:861–874

123

concentration, which were within experimental uncer-

tainty, were observed between the influent and effluent

concentrations. The lack of appreciable NO3- loss in

control experiments was consistent with the assertion that

C limited NO3- loss in our soil.

Rates of Nitrous Oxide Production

Similar to NO3- loss, N2O production rates (Eq. 2) were

greatest for citric acid experiments at all concentrations,

followed by the alginic acid and SRDOC experiments

(Table 2). In experiments with citric acid, mean N2O

production rates increased linearly with increased citric

acid concentration (Fig. 2b; R2 = 0.92). In contrast, N2O

production rates in alginic acid and SRDOC treatments

were approximately an order of magnitude lower than the

rate at the corresponding citric acid concentration for

experiments with 8 and 16 mg C L-1. Although all N2O

production rates were above the LOD, production rates for

alginic acid and SRDOC experiments were statistically

different from controls only in experiments with

16 mg C L-1. These results suggest that alginic acid and

SRDOC stimulate to N2O production in our system only

when present at high concentration, consistent with low

NO3- loss rates. However, rates of NO3 loss were between

100- to 250-fold greater that N2O production rates, indi-

cating that only a small portion of NO3- lost was converted

to N2O.

Quantification of nir and 16s rDNA Genes

The 16s rDNA, nirK, and nirS abundances in the unit of

copy number g-1 soil (Table 3) were grouped by C type

and plotted as a function of concentration in Fig. 3. There

was a clear trend of increasing 16s rDNA gene copy

numbers (8 ± 3 9 107 to 1.8 ± 0.3 9 108 copy g-1 soil)

0

0.2

0.4

0.6

0.8

1.0N

O3-

loss

rat

e

Carbon concentration (mg C L-1)

Citric acid

Alginic acid

SRDOC

0

2.0

4.0

6.0

N2O

pro

duct

ion

rate

(mg

N h

r-1 g

-1)

× 10

-6 Citric acid

Alginic acid

SRDOC

0 4 8 12 16

a

Carbon concentration (mg C L-1)0 4 8 12 16

b

(mg

N h

r-1 g

-1)

× 10

-3

R2 = 0.98

R2 = 0.92

Fig. 2 Average rates of a NO3- loss and b NO2 production rates as a

function of carbon type and concentration. Lines represent a least-

squares fit of a linear model to the citric acid data. For both NO3- loss

and N2O production rates, uncertainty is reported as twice the

standard deviation (±1 standard deviation) of three replicates of

experiments

nirK

copy

num

ber

(×10

8 )

0 mg C L-1

4 mg C L-1

8 mg C L-1

16 mg C L-1

nirS

copy

num

ber

(×10

8)

0 mg C L-1

4 mg C L-1

8 mg C L-1

16 mg C L-1

0

0.5

1.0

1.5

2.0

2.5

Citric acid

16 r

DN

A c

opy

num

ber

(×10

8)

Carbon type

0 mg C L-1

4 mg C L-1

8 mg C L-1

16 mg C L-1

a

SWDOCAlginic acid

b

c

Citric acid

Carbon type

SWDOCAlginic acid

Citric acid

Carbon type

SWDOCAlginic acid

0

0.5

1.0

1.5

2.0

2.5

0

0.5

1.0

1.5

2.0

2.5

a

b

b

c

ab b b

a b b b

a

bb b

a b b b aaaa

aa a a

aaaaa

bb

b

Fig. 3 The a nirK b nirS, and c 16s rDNA copy numbers as a

function of carbon concentration (mg C L-1) for different carbon

types. Uncertainty is reported as twice the standard deviation

(±1 standard deviation) of nine amplifications at each condition.

Within the same treatment, bars with the same letter above them are

not significantly different (P [ 0.05) based on a LSD test

Environmental Management (2012) 50:861–874 867

123

with increasing C concentration in the citric acid treatment.

Experiments containing alginic acid and SRDOC contained

higher 16s rDNA gene copy numbers than control experi-

ments; however, gene copy numbers were lower than in the

corresponding experiments with citric acid, and did not

vary with increasing C concentrations.

Soil samples in the citric acid treatment contained more

nir gene copies than in control experiments or corre-

sponding experiments with the other C sources (Fig. 3a, b).

However, although all citric acid experiments contained

higher nir copy numbers than the control, there was no

trend in the average nirK or nirS gene copy numbers with

increasing C concentrations. Experiments with alginic acid

had higher nirS numbers than control, but nirK gene copy

numbers were not significantly different from the control.

Similarly, experiments with SRDOC did not produce an

increase in nir genes copy number from the control. Soil

samples with alginic acid and SRDOC contained fewer nir

gene copies per gram than corresponding experiments with

citric acid for all C concentrations.

Discussion

Effect of Organic Carbon on Rates of Nitrate Loss

and Nitrous Oxide Production

Organic C availability often limits denitrification in

groundwater, riparian buffers, and other types of wetlands

(Greenan and others 2006; Rivett and others 2008), and has

been suggested to limit activity at in soils at CEFS, the site

from which these samples were obtained (Knies 2009).

However, C type, as well as concentration, may regulate

organic C availability, and thus the denitrification rate.

Denitrification has been shown to correlate more strongly

with readily decomposable soil organic matter than with

total organic C (Beauchamp and others 1980; Bijay-Singh

and Whitehead 1988; Burford and Bremner 1975; David-

son and others 1987; Hill and Cardaci 2004; Myrold

and Tiedje 1985; Stanford and others 1975; McCarty and

Bremner 1992; Pintar and Lobnik 2005; Hernandez and

Mitsch 2007). The nature of organic C may also be

responsible for ecosystem-scale differences in denitrifica-

tion (Ward and others 2009).

In our experiments, citric acid was most effective in

promoting NO3- loss and N2O production. Citric acid, a

LMMOA and intermediate in common metabolic cycles, is

likely to be readily available to organisms, and thus pro-

moted NO3- reduction in our experiments. The release of

LMMOAs, such as citric acid, malic acid, and acetic acid,

from root exudation and cell lysis has been well docu-

mented in surface soil and the rhizosphere (Jones 1998;

Jones and others 2003). These molecules are rapidly

mineralized in soils (van Hees and others 2005), and thus

are a good substrate for microbial growth and the con-

sumption of electron acceptors. Our results suggested that

LMMOAs in soil organic C may be tightly coupled to

NO3- loss, in agreement with results from previous labo-

ratory (Henry and others 2008) and field studies (Stow and

others 2005; Sirivedhin and Gray 2006; Bernhardt and

Likens 2002).

Several similar column studies provide a basis of com-

parison for the rates of NO3- loss derived from our study.

Pavel et al. (1996) utilized surface soils with high DOC

concentrations (9.3 % and 7.8 %) without additional C

addition, but used a similar continuous-flow column

experimental approach. In situations where transport pro-

cesses did not control the reaction, their rates were

approximately 10-4 to 10-5 mg N h-1 g-1 soil. Knies

(2009) employed a similar methodology and concentration

range as the current study, but used D-glucose as C source;

NO3- loss rates also varied between 10-4 to 10-5

mg N h-1 g-1 soil, and generally increased with increas-

ing C concentration. These results thus broadly agree with

our measured rates of NO3- loss (Table 2).

Alginic acid had a small effect NO3- loss rates and only

promoted N2O production at high concentration in our

experiments. Many authors have suggested that soluble C

sources are key drivers of microbial activity, although not

all soluble C is equally bioavailable (Boyer and Groffman

1996; Cook and Allan 1992). The generally accepted

dogma is that high molecular mass organic molecules are

more recalcitrant than low molecular mass organic mole-

cules (Tate 1987), which could be a possible explanation

for the low NO3- loss and N2O production rates in our

alginic acid column experiments. This observation is sup-

ported by previous work that demonstrated that glucose

enhanced the denitrification potential more effectively than

the biopolymer cellulose or grass litter (Dendooven and

others 1996).

Our results indicated that SRDOC did not affect NO3-

loss rates and only promoted N2O production at high

concentration in our experiments. This result was consis-

tent with studies that have shown denitrification to be

limited even in the presence of abundant DOC (Starr and

Gillham 1993), perhaps because of the nature of the C type.

Previous results have also demonstrated that denitrification

in riparian buffer soils is highest in surface horizons and

depends on organic matter content (Groffman and others

1992; Lowrance 1992; Groffman and others 1991; Willems

and others 1997); however, surface horizons have high C

concentrations that may contain more labile types of C in

addition to humic substances (Thurman 1985; Tate 1987).

One possible reason for the low reactivity of the SRDOC is

that humic substances tend to be recalcitrant, and thus have

long environmental residence times and degrade slowly via

868 Environmental Management (2012) 50:861–874

123

microbial processes (Thurman 1985). Additionally,

SRDOC sample has a high C/N ratio (48:1), which may not

be optimal for supporting microbial activity. Van Mooy

and others (2002) found that denitrifying microbes pref-

erentially consumed N-rich organic C when given suffi-

cient C supply. Also, Groffman and others (1991) found

higher denitrification potential in grassy vegetated filter

strips as compared to forest strips, which was attributed to

lower C/N ratio (\10) in the grassy buffer.

Although our results indicated low rates of NO3- loss

and N2O production for alginic acid and SRDOC, the

molecules in fact may promote N transformations in the

field. The residence time in our columns (ca. 2.8 h) was

short when compared to the typical time required for

groundwater to pass through a riparian buffer, wetlands, or

the hyporheic zone. Thus, if these molecules are abundant

in a specific environment, they may appreciably promote

denitrification, albeit at a slower rate, as long as other

conditions are favorable and the residence time of the

groundwater is sufficient (Ocampo and others 2006). This

idea is also supported by the observed increase in nirS gene

copy number upon addition of alginic acid, which suggests

some stimulation of N cycle bacteria in our soil.

It is also important to note the limitations of the

chemical measurements described in the present study.

Measurements of N transformations, even in carefully

controlled experimental conditions, are complicated

because myriad processes are possible. In our anaerobic

columns, where the major electron donor and acceptor

were organic C and NO3-, respectively, it is reasonable to

assume that denitrification, a process that produces gaseous

nitrogen oxides and N2, was occurring in our system.

However, several different factors convoluted the inter-

pretation of our kinetic data. First, it should be noted that

the mass of NO3- lost in our system greatly exceeded the

mass of N2O produced for all measurements, consistent

with previous conceptual models of denitrification (Fire-

stone and Davidson 1989). However, N2O may have been

produced by other processes [e.g., nitrification (Bremner

and Blackmer 1978)], and thus other N transformation

processes may have affected our N2O production rates.

Second, N2, the major production of denitrification, was

not measured due to the difficulties in its quantification

(Groffman and others 2006). We were thus unable to

complete a full mass balance for N in our system. This

limited our ability to relate influent NO3-concentration to

the concentration of N compounds in the effluent. Finally,

although the N content of our soil is low (0.02 %), it was

not possible to conclusively exclude the possibility that N

mineralization and nitrification may result in internal pro-

duction of NO3- in our columns, which could convolute

mass balance approaches and may have affected our

measured rates. The possible contributions of other N

transformation processes thus limit the strict interpretation

of the kinetic measurements in these columns to net rates of

NO3- loss and N2O production.

Effect of Organic Carbon Gene Abundances

Microbial communities may be altered by C amendments

that select for populations that are most competitive in

terms of growth rates and their ability to utilize nutrients

(Drenovsky and others 2004). Our study found an increase

in nirK gene abundance in citric acid experiments as

compared to untreated controls, but no change for systems

with alginic acid or SRDOC. Henry and others (2004)

observed a four-fold increase in nirK gene copy number in

soil amended with a unspecified organic C source. Simi-

larly, the nirK abundance correlated with quantity of soil

organic C in a receding glacier foreland, although little

information was provided regarding the type of C (Kan-

deler and others 2006). Taken together, these observations

suggest that the type and concentration of C are key factors

in controlling the denitrifying bacteria community.

16s rDNA copy numbers have been commonly used to

estimate the bacterial population in soils. Compared to

previous studies, which found the 16s rDNA copy numbers

in agricultural soils to be ca. 6 9 107 to 1 9 108 copy g-1

soil (Dandie and others 2008; Miller and others 2009), our

16s rDNA gene abundances were somewhat low (Table 1),

consistent with the previous assertion that low concentra-

tions of organic C may be limiting biological activity in

CEFS soils (Knies 2009). In our soil, 16s rDNA gene

abundances increased upon addition of all types of C,

suggesting stimulation of the overall microbial community

by C addition.

Copy numbers of nir genes, which potentially have a

functional relationship with N reduction rates, may also be

compared to prior results in observed in soils. Our results for

nirK in trials without C addition (5 9 106 copy g-1 soil)

were in rough agreement with previous studies of environ-

mental samples that have reported 105 to 106 copy g-1 soil

of nirK in sandy soils (Henry and others 2004), 5 9 104

to 107 copy g-1 soil in a subsoil of a fir plantation (Levy-

Booth and Winder 2010), 4.1-7.3 9 107 copy g-1 soil in a

agricultural field, 107 copy g-1 soil in rice paddy soil

(Yoshida and others 2009), and 107 copy g-1 soil of glacial

soils (Kandeler and others 2006). Our nirS gene abundance

for soil without C addition (2 9 106 copy g-1 soil) agreed

well with what has been reported for other unamended

environmental samples, including estuarine sediments (105

to 106 copy g-1 soil) (Smith and others 2006), a subsoil of a

fir plantation (3 9 105 copy g-1 soil) (Levy-Booth and

Winder 2010), rice paddy soil (106 copy g-1 soil) (Yoshida

and others 2009) and soil of glacial soils (5 9 106 to 1 9 108

copy g-1) (Kandeler and others 2006).

Environmental Management (2012) 50:861–874 869

123

Bacterial Community Composition

Table 4 shows the ratios of the genes in soils from our

column experiments. Changes in the ratios provide infor-

mation on how community composition changes with C

addition. The ratio of nirK to 16s rDNA varies from 0.50-

0.88, a somewhat surprising result that reflects the low 16s

rDNA copy number measured in our soils. Although some

studies have found a significantly smaller ratio (Kandeler

and others 2006; Henry and others 2008), Daidie and others

(2008) found similar ratios of nirK:16s rDNA in an agri-

cultural field soil. nirS genes were generally less abundant

than nirK in our soil, with ratios of nirK to nirS ranging

from 1.3 to 2.4 (Table 4). The ratio of these genes in lit-

erature studies of soils is highly variable (Bothe and others

2000; Yoshida and others 2009, 2010; Kandeler and others

2006; Levy-Booth and Winder 2010; Henry and others

2008; Dandie and others 2011), suggesting that local fac-

tors may control which gene is more abundant. However,

the relative consistency of ratios of nirK:nirS in our

experiments indicated that the denitrifier community

composition is relatively stable regardless of the C type and

concentration, in accordance with a previous study that

noted a correlation between nirK and nirS gene abundance

(Levy-Booth and Winder 2010). Although multiple gene

copies may be present in a single cell [Philipott (2002)

reported that only one copy of nirK is present in a bacte-

rium] and the primer bias cannot be excluded, our qPCR

results in control experiments suggest that the nirK-har-

boring denitrifiers might be more abundant than nirS-har-

boring denitrifiers in our riparian buffer soil.

Gene Abundances and Rates of Nitrate Loss

and Nitrous Oxide Production

Although nir genes are responsible for NO2-, not NO3

- or

NO, reduction, it is reasonable to postulate that the rates of

these processes are strongly correlated, especially given the

previous observation that NO is an unstable intermediate in

denitrification (Firestone and Davidson 1989) and our low

observed concentrations of NO2- in the effluent. In our study,

the nirK and nirS gene copy numbers trended with increasing

rates of NO3- loss or N2O production (Fig. 4a–d); however,

there was not a linear correlation between either nir gene

abundance and these rates. In contrast, the 16s rDNA gene

copy number (Fig. 4e, f) correlated with both NO3- loss rate

(R2 = 0.85) and N2O production rate (R2 = 0.72). Overall,

our chemical and biological measurements were consistent

with previous work that noted an increase in both the deni-

trification rate, and the abundance of denitrifying and total

bacteria (as determined by most probable number analysis) in

soils amended with high concentrations of glucose and NO3-

(Ardakani and others 1975; Bowman and Focht 1974). It is

important to remember other biological factors may affect the

relationship between copy gene abundance and denitrifica-

tion rate. Most notably, elevated metabolic activity, which

may be measured by nir gene expression (mRNA quantifi-

cation) instead of gene abundance, could also result in

increased NO3- loss and N2O production rates.

Henry and others (2008) and Mounier and others (2004)

observed increases in denitrification rates but only small

changes in denitrifying bacterial community composition in

agricultural soils amended with root exudates (a mixture of

sugars, organic acids, and amino acids, which are somewhat

analogous to citric acid) and corn mucilage (a complicated

mixture of polysaccharides, which are structurally related to

alginic acid). In the context of our findings, their results are

somewhat surprising in that their experiments are somewhat

analogous to our experiments with citrate, which resulted in

an increase in NO3- loss rate, N2O production rate, and in nir

gene copy number, and with alginate, which resulted in small

changes to NO3- loss rates and nirS gene copy numbers.

Although some of these differences may result from differ-

ences in experimental conditions (e.g., silty clay loam vs.

sand, macrocosm vs. continuous flow column experiment,

concentration of NO3- and organic C, etc.), the results also

underscore the potential importance of specific molecules in

preferentially promoting specific microbial processes.

It should be noted that, as in any PCR-based study, there

are artifacts associated with amplification process. The use of

specific primers targeting specific genetic sequences may not

allow the quantification of all the target functional genes in a

population. The possible presence of multiple gene copy

numbers in an organism may distort the view of community

composition. In addition, PCR may amplify genes that are in

dead or dormant organisms. However, because we measured

changes with C addition from an unamended control system,

the significance of the last point is minimized. It also is

possible that storage before amplification may have affected

the microbial community.

Table 4 Ratios of the NO2- reduction genes nirK and nirS copy

numbers to the 16s rDNA copy number

Type Concentration

(mg C L-1)

nirK/16S nirS/16S nirK/nirS

Citric Acid 0 0.83 0.33 2.5

4 0.88 0.38 2.3

8 0.58 0.33 1.8

16 0.50 0.22 2.3

Alginic Acid 4 0.85 0.35 2.4

8 0.85 0.48 1.8

16 0.67 0.40 1.7

SRDOC 4 0.86 0.43 2.0

8 0.86 0.64 1.3

16 0.77 0.47 1.7

870 Environmental Management (2012) 50:861–874

123

The results of this study indicate that structure of soluble

C may be critical in regulation of soil N transformations,

thus highlighting the potential importance of vegetation in

promoting denitrification in riparian buffers and other ter-

restrial landscapes. Although a recent meta-analysis did not

identify vegetation as a key factor for controlling subsur-

face N removal in riparian buffers, the effect may have

been masked in such an analysis by other soil biogeo-

chemical factors (Mayer and others 2007). Metabolites,

such as citric acid and other LMMOAs, are often leaked

from roots, accounting for up to one-fifth of the C fixed by

plants (Nguyen 2003), and thus are often present in high

concentrations in the rhizosphere (Jones 1998; Curl and

Truelove 1986; Jones and others 2003) where they may

affect the microbial ecology of denitrifying bacteria

(Philippot and others 2007). In some environments,

LMMOA and other small molecules may account for up to

30 % of organic C (van Hees and others 2005). These

molecules, which have short environmental residence

times, are critical drivers of many environmental processes,

and may also be important connections between the C and

N cycles.

Conclusions

The rates of N transformations in soil may vary with

organic C concentrations and type. When sufficient NO3-

is present in system, availability of liable organic C may

limit denitrification under anaerobic conditions. Results

from molecular biology approaches indicated that specific

components of soil organic matter may have differing

effects on the rates of N transformations by influencing

gene abundances in riparian buffer soil. Citric acid, a labile

organic metabolite, exhibited significant enhancement of N

transformation rates and gene abundances; however,

increased concentrations of alginic acid and SRDOC either

did not affect or had a small positive effect on NO3- loss

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

nirK

copy

num

ber

(×10

8 )

0

0.2

0.4

0.6

0.8

nirS

copy

num

ber

(×10

8 )

a b

c

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

nirK

copy

num

ber

(×10

8)

0 1 2 3 4 50 0.2 0.60.4 0.8

0 0.2 0.60.4 0.80

0.2

0.4

0.6

0.8

nirS

copy

num

ber

(×10

8 )

0 1 2 3 4 5

d

0

0.5

1.0

1.5

2.0

2.5

16s

rDN

Aco

pynu

mbe

r ( ×

10 )

0 1 2 3 4 50 0.2 0.60.4 0.8

16s

rDN

Aco

pynu

mbe

r ( ×

10 )

0

0.5

1.0

1.5

2.0

2.5

N2O production rate (mg N hr-1 g-1) × 10-6

e fN2O production rate (mg N hr-1 g-1) × 10-6

N2O production rate (mg N hr-1 g-1) × 10-6

NO3- loss rate (mg N hr-1 g-1) × 10-3

NO3- loss rate (mg N hr-1 g-1) × 10-3

NO3- loss rate (mg N hr-1 g-1) × 10-3

R2 = 0.85 R2 = 0.72

8 8

Fig. 4 NO3- loss rates versus

a nirK, c nirS, and e 16s rDNA

gene copy numbers;

N2Oproduction rates vs. b nirK,

d nirS, and f 16s rDNA gene

copy numbers. Lines represent a

least-squares fit of a linear

model to the 16s rDNA data.

Legend: squares alginic acid,

triangle SWDOC, diamondscitric acid

Environmental Management (2012) 50:861–874 871

123

rate, N2O production, and the copy number of functional

genes. The ratio of nirK:nirS genes in amended soils was

relatively constant, suggesting reasonable stability in the

denitrifying community structure. Longer residence time

experiments with complex or recalcitrant C sources may

help to further refine and quantify their effect on N trans-

formation rates and microbial community dynamics.

Additionally, more field studies that relate differing types

of C to functional gene copy numbers or microbial com-

munity dynamics may help to further refine the relation-

ships between N transformations, biological activity, and C

availability.

Acknowledgments We thank Lauren Saal, Emily Dell, Guillemo

Ramirez, Sara Knies, Brad Robinson, Chris Ashwell, Greg Dick,

Suzanna Brauer, Jeff White, and John Walker. This work was sup-

ported by NC Department of Environment & Natural Resources,

Conservation Reserve Enhancement Program. Lin Wu thanks the

Society of Wetland Scientists for a student research grant.

References

Ardakani MS, Belser LW, McLaren AD (1975) Reduction of nitrate

in a soil column during continuous flow. Soil Science Society of

America Journal 39:290–294

Barnhill WL, Goodwin J, Bostian MR, McLoda NA, Leishman GW,

Scanu RJ (1974) Soil survey of Wayne County, North Carolina.

Soil Conservation Service, Washington, DC

Beauchamp EG, Gale C, Yeomans JC (1980) Organic matter

availability for denitrification in soils of different textures and

drainage classes. Communications in Soil Science and Plant

Analysis 11:1221–1233

Bernhardt ES, Likens GE (2002) Dissolved organic carbon enrich-

ment alters nitrogen dynamics in a forest stream. Ecology

83:1689–1700

Bijay-Singh RydenJC, Whitehead DC (1988) Some relationships

between denitrification potential and fractions of organic carbon

in air-dried and field-moist soils. Soil Biology & Biochemistry

20:737–741

Bothe H, Jost G, Schloter M, Ward BB, Witzel KP (2000) Molecular

analysis of ammonia oxidation and denitrification in natural

environments. FEMS Microbiology Reviews 24:673–690

Bowman RA, Focht DD (1974) Influence of glucose and nitrate

concentrations upon denitrification rates in sandy soils. Soil

Biology & Biochemistry 6:297–301

Boyer JN, Groffman PM (1996) Bioavailability of water extractable

organic carbon fractions in forest and agricultural soil profiles.

Soil Biology & Biochemistry 28:783–790

Bradley PM, Fernandez M, Chapelle FH (1992) Carbon limitation of

denitrification rates in an anaerobic groundwater system. Envi-

ronmental Science and Technology 26:2377–2381

Bremner JM, Blackmer AM (1978) Nitrous oxide: emission from

soils during nitrification of fertilizer nitrogen. Science 199:

295–296

Burford JR, Bremner JM (1975) Relationships between denitrification

capacities of soils and total, water soluble and readily decom-

posable soil organic-matter. Soil Biology & Biochemistry 7:

389–394

Cook BD, Allan DL (1992) Dissolved organic carbon in old field

soils- compositional changes during the biodegredation of soil

organic-matter. Soil Biology & Biochemistry 24:595–600

Curl EA, Truelove B (1986) The rhizosphere. Springer, New York

Dandie CE, Burton DL, Zebarth BJ, Henderson SL, Trevors JT,

Goyer C (2008) Changes in bacterial denitrifier community

abundance over time in an agricultural field and their relation-

ship with denitrification activity. Applied and Environmental

Microbiology 74:5997–6005

Dandie CE, Wertz S, Leclair CL, Goyer C, Burton DL, patten CL,

Zebarth BJ, Trevors JT (2011) Abundance, diversity and

functional gene expression of dentrifier comminties in adjacent

riparian and agricultural zones. FEMS Microbiology Ecology

77:69–82

Davidson EA, Galloway LF, Strand MK (1987) Assessing available

carbon: comparison of techniques across selected forest soils.

Communications in Soil Science and Plant Analysis 18:45–64

Dendooven L, Splatt P, Anderson JM (1996) Denitrification in

permanent pasture soil as affected by different forms of C

substrate. Soil Biology & Biochemistry 28:141–149

Dosskey MG, Vidon P, Gurwick NP, Allan CJ, Duval TP, Lowrance

R (2010) The role of riparian vegetation in protecting and

improving chemical water quality in streams. Journal of the

American Water Resources Association 46:261–277

Drenovsky RE, Vo D, Graham KJ, Scow KM (2004) Soil water

content and organic carbon availability are major determinants

of soil microbial community composition. Microbial Ecology

48:424–430

Dukes MD (2000) Effect of riparian buffers and controlled drainage

on shallow groundwater quality in the North Carolina Middle

Coastal Plain. Doctoral Dissertation, North Carolina State

University, Raleigh

Dukes MD, Evans RO, Gilliam JW, Kunickis SH (2002) Effect of

riparian buffer width and vegetation type on shallow groundwa-

ter quality in the Middle Coastal Plain of North Carolina.

Transactions of the ASAE 45:327–336

Firestone MK, Davidson EA (1989) Microbiological basis of NO and

N2O production and consumption in soil. In: Andreae MO,

Schimel DS (eds) Exchange of trace gases between terrestrial

ecosystems and the atmosphere, vol 47. Life Sciences Research

Report. Wiley, Hoboken, NJ, pp 7–21

Gee GW, Or D (2002) Particle-size analysis. In: Dane JH, Topp GC

(eds) Methods of soil analysis. Part 4. Physical methods. Soil

Science Society of America, Madison, WI

Greenan CM, Moorman TB, Kaspar TC, Parkin TB, Jaynes DB

(2006) Comparing carbon substrates for denitrification of

subsurface drainage water. Journal of Environmental Quality

35:824–829

Groffman PM, Axelrod EA, Lemunyon JL, Sullivan WM (1991)

Denitrification in grass and forest vegitated filter strips. Journal

of Environmental Quality 20:671–674

Groffman PM, Gold AJ, Simmons RC (1992) Nitrate dynamics in

riparian forests-microbial studies. Journal of Environmental

Quality 21:666–671

Groffman PM, Altabet MA, Bohlke JK, Butterbach-Bahl K, David

MB, Firestone MK, Giblin AE, Kana TM, Nielsen LP, Voytek

MA (2006) Methods for measuring denitrification: diverse

approaches to a difficult problem. Ecological Applications

16:2091–2122

Hedin LO, von Fischer JC, Ostrom NE, Kennedy BP, Brown MG,

Robertson GP (1998) Thermodynamic constraints on nitrogen

transformations and other biogeochemical processes at soil-

stream interfaces. Ecology 79:684–703

Henry S, Baudoin E, Lopez-Gutierrez JC, Martin-Laurent F,

Baumann A, Philippot L (2004) Quantification of denitrifying

bacteria in soils by nirK gene targeted real-time PCR. Journal of

Microbiological Methods 59:327–335

Henry S, Texier S, Hallet S, Bru D, Dambreville C, Cheneby D, Bizouard

F, Germon JC, Philippot L (2008) Disentangling the rhizosphere

872 Environmental Management (2012) 50:861–874

123

effect on nitrate reducers and denitrifiers: insight into the role of root

exudates. Environmental Microbiology 10:3082–3092

Hernandez ME, Mitsch WJ (2007) Denitrification potential and

organic matter as affected by vegetation community, wetland

age, and plant introduction in created wetlands. Journal of

Environmental Quality 36:333–342

Hill AR (1996) Nitrate removal in stream riparian zones. Journal of

Environmental Quality 25:743–755

Hill AR, Cardaci M (2004) Denitrification and organic carbon

availability in riparian wetland soils and subsurface sediments.

Soil Science Society of America Journal 68:320–325

Hill AR, Devito KJ, Campagnolo S, Sanmugadas K (2000) Subsur-

face denitrification in a forest riparian zone: interactions between

hydrology and supplies of nitrate and organic carbon. Biogeo-

chemistry 51:193–223

Jones DL (1998) Organic acids in the rhizosphere: a critical review.

Plant and Soil 205:25–44

Jones DL, Dennis PG, Owen AG, van Hees PAW (2003) Organic acid

behavior in soils: misconceptions and knowledge gaps. Plant and

Soil 248:31–41

Kandeler E, Deiglmayr K, Tscherko D, Bru D, Philippot L (2006)

Abundance of narG, nirS, nirK, and nosZ genes of denitrifying

bacteria during primary successions of a glacier foreland.

Applied and Environmental Microbiology 72:5957–5962

King SE (2006) Riparian buffer effectiveness in removing ground-

water nitrate as influenced by vegetative type. Master’s Thesis,

North Carolina State University, Raleigh

Knies SV (2009) Riparian buffer effectiveness at removal of NO3-N

from groundwater in the middle Coastal Plain of North Carolina.

Master’s Thesis, North Carolina State University, Raleigh

Knowles R (1982) Denitrification. Microbiological Reviews 46:43–70

Levy-Booth DJ, Winder RS (2010) Quantification of nitrogen

reductase and nitrite reductase genes in soil of thinned and

clear-cut Douglas-Fir stands by using real-time PCR. Applied

and Environmental Microbiology 76:7116–7125

Lowrance R (1992) Groundwater nitrate and denitrification in a

coastal plain riparian forest. Journal of Environmental Quality

21:401–405

Lowrance R, Altier LS, Newbold JD, Schnabel RR, Groffman PM,

Denver JM, Correll DL, Gilliam JW, Robinson JL, Brinsfield

RB, Staver KW, Lucas W, Todd AH (1997) Water quality

functions of riparian forest buffers in Chesapeake Bay water-

sheds. Environmental Management 21:687–712

Mayer PM, Reynolds SK, McCutchen MD, Canfield TJ (2007) Meta-

analysis of nitrogen removal in riparian buffers. Journal of

Environmental Quality 36:1172–1180

McCarty GW, Bremner JM (1992) Availability of organic carbon for

denitrification of nitrate in subsoils. Biology and Fertility of

Soils 14:219–222

Miller MN, Zebarth BJ, Dandie CE, Burton DL, Goyer C, Trevors JT

(2009) Denitrifier community dynamics in soil aggregates under

permanent grassland and arable cropping systems. Soil Science

Society of America Journal 73:1843–1851

Miller MN, Dandie CE, Zebarth BJ, Burton DL, Goyer C, Trevors JT

(2012) Influence of carbon amendments on soil denitrifier

abundance in soil microcosms. Geoderma 170:48–55

Mounier E, Hallet S, Cheneby D, Benizri E, Gruet Y, Nguyen C,

Piutti S, Robin C, Slezack-Deschaumes S, Martin-Laurent F,

Germon JC, Philippot L (2004) Influence of maize mucilage on

the diversity and activity of the denitrifying community.

Environmental Microbiology 6:301–312

Myrold DD, Tiedje JM (1985) Establishment of denitrification

capacity in soil: effects of carbon, nitrate and moisture. Soil

Biology & Biochemistry 17:819–822

Nguyen C (2003) Rhizodeposition of organic C by plants: mecha-

nisms and controls. Agronomie 23:375–396

Nogales B, Timmis KN, Nedwell DB, Osborn AM (2002) Detection and

diversity of expressed denitrification genes in estuarine sediments

after reverse transcription-PCR amplification from mRNA.

Applied and Environmental Microbiology 68:5017–5025

Nommik H (1956) Denitrification in soil. Acta Agriculturæ Scandi-

navica 6:195–228

Ocampo CJ, Oldham CE, Sivapalan M (2006) Nitrate attenuation in

agricultural catchments: shifting balances between transport and

reaction. Water Resources Research 42:W01408

Osborne LL, Kovacic DA (1993) Riparian vegetated buffer strips in

water-quality restoration and stream management. Freshwater

Biology 29:243–258

Pavel EW, Reneau RB, Berry DF, Smith EP, Mostaghimi S (1996)

Denitrification potential of nontidal riparian wetland soils in the

Virginia coastal plain. Water Research 30:2798–2804

Perry TD, Duckworth OW, McNamara CJ, Martin ST, Mitchell R

(2004) The effects of the biologically produced polymer alginic

acid on macroscopic and microscopic calcite dissolution rates.

Environmental Science and Technology 38:3040–3046

Perry TD, Cygan RT, Mitchell R (2006) Molecular models of alginic

acid: interactions with calcium ions and calcite surfaces.

Geochimica et Cosmochimica Acta 70:3508–3532

Philippot L (2002) Denitrifying genes in bacterial and archaeal

genomes. Biochimica et Biophysica Acta (BBA) - Gene Structure

and Expression 1577:355–376

Philippot L, Hallin S, Schloter M (2007) Ecology of denitrifying

prokaryotes in agricultural soil. Advances in Agronomy 96:

249–305

Pintar M, Lobnik F (2005) The impact of nitrate and glucose availability

on the denitrification at different soil depths. Fresenius Environ-

mental Bulletin 14:514–519

Rivett MO, Buss SR, Morgan P, Smith JWN, Bemment CD (2008)

Nitrate attenuation in groundwater: a review of biogeochemical

controlling processes. Water Research 42:4215–4232

Sirivedhin T, Gray KA (2006) Factors affecting denitrification rates in

experimental wetlands: field and laboratory studies. Ecological

Engineering 26:167–181

Smith TA, Osmond DL, Gilliam JW (2006) Riparian buffer width and

nitrate removal in a lagoon-effluent irrigated agricuttural area.

Journal of Soil and Water Conservation 61:273–281

Stanford G, Vanderpol RA, Dzienia S (1975) Denitrification rates in

relation to total and extractable soil carbon. Soil Science Society

of America Journal 39:284–289

Starr RC, Gillham RW (1993) Denitrification and organic carbon

availability in two aquifers. Ground Water 31:934–947

Stow CA, Walker JT, Cardoch L, Spence P, Stow CA (2005) N2O

emissions from streams in the Neuse River watershed, North

Carolina. Environmental Science and Technology 39:6999–7004

Stumm W, Morgan JJ (1996) Aquatic chemistry, 3rd edn. Wiley, New

York

Tate RL (1987) Soil organic matter: biological and ecological effects.

Wiley, New York

Throback IN, Enwall K, Jarvis A, Hallin S (2004) Reassessing PCR

primers targeting nirS, nirK and nosZ genes for community

surveys of denitrifying bacteria with DGGE. FEMS Microbiol-

ogy Ecology 49:401–417

Thurman EM (1985) Organic geochemistry of natural waters. Kluwer,

Boston

van Hees PAW, Jones DL, Finlay R, Godbold DL, Lundstom US

(2005) The carbon we do not see: the impact of low molecular

weight compounds on carbon dynamics and respiration in forest

soils: a review. Soil Biology & Biochemistry 37:1–13

Van Mooy BAS, Keil RG, Devol AH (2002) Impact of suboxia on

sinking particulate organic carbon: enhanced carbon flux and

preferential degradation of amino acids via denitrification.

Geochimica et Cosmochimica Acta 66:457–465

Environmental Management (2012) 50:861–874 873

123

Wallenstein MD, Myrold DD, Firestone M, Voytek M (2006)

Environmental controls on denitrifying communities and deni-

trification rates: insights from molecular methods. Ecological

Applications 16:2143–2152

Ward BB, Devol AH, Rich JJ, Chang BX, Bulow SE, Naik H,

Pratihary A, Jayakumar A (2009) Denitrification as the dominant

nitrogen loss process in the Arabian Sea. Nature 461:78–81

Willems HPL, Rotelli MD, Berry DF, Smith EP, Reneau RB,

Mostaghimi S (1997) Nitrate removal in riparian wetland soils:

effects of flow rate, temperature, nitrate concentration and soil

depth. Water Research 31:841–849

Yoshida M, Ishii S, Otsuka S, Senoo K (2009) Temporal shifts in

diversity and quantity of nirS and nirK in a rice paddy field soil.

Soil Biology & Biochemistry 41:2044–2051

Yoshida M, Ishii S, Otsuka S, Senoo K (2010) nirK-harboring

denitrifiers are more responsive to denitrification inducing

conditions in rice paddy soil than nirS-harboring bacteria.

Microbes and Environments 25:45–48

874 Environmental Management (2012) 50:861–874

123