:7... ...« Jurariv

.

p5...” a. .:. .3 7;...

.2 Jr: . .:.

.

... t n}...

[.:.

 

 

; 1858's
. I a
9007

This is to certify that the
dissertation entitled

THE DIVERSITY OF DISSIMILATORY NITRATE REDUCERS
IN AN AGROECOSYSTEM

presented by

KRISTIN MICHELLE HUIZINGA

has been accepted towards fulfillment

of the requirements for the
Microbiology & Molecular
PhD degree in Genetics

 

 

 

 

VMProfessor’s Signature
flmc “.374 a5, 2 0%
d 1
Date

AflflLkanAMhmnNoAdhnEhmflOnmanmHnummbn

 

LIBRARY
Michigan State
Unlyersity

 

 

 

00-.-.-----o-o--o--o----:---:--o-IC-----o------------u-.-.-.-I-u--------I-o-o-c-o-u-coc-o-o-o-u-c----.-c.

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DaleDue.indd-p.1

THE DIVERSITY OF DISSIMILATORY NITRATE REDUCERS IN AN
AGROECOSYSTEM

by

Kristin Michelle Huizinga

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Microbiology & Molecular Genetics

2006

ABSTRACT

The Diversity of Dissimilatory Nitrate Reducers in an Agroecosystem

by

Kristin Michelle Huizinga

Microbial communities are essential to nutrient cycles in agricultural soils, since
their activity helps regulate the availability of nutrients to crops. Despite this important
role, factors affecting microbial community structure are just beginning to be uncovered.
As well as identifying factors affecting microbial communities, it is important to
determine the links between particular communities and their ecosystem functions.
Therefore, the focus of the work presented is on microbial communities known to be
involved in soil nitrogen cycling (denitrifiers) or with the potential for involvement
(Planctomycetales).

Statistical modeling determined how much of the variability in N20 and two other
greenhouse gas fluxes (C02 and CH4) could be explained from treatments varying in
history of agriculture. The amount of explained variation was generally low, with C02
flux having the greatest amount of explainable variation. It was concluded that the high
amount of explainable variability in C02 flux was due to the fact that it is a byproduct of
heterotrophic metabolism in soil, making its production responsive to any factors
affecting microbial metabolism. Nitrous oxide and CH4, conversely, are produced or
consumed by microorganisms with specialized metabolisms, causing environmental

variables alone to be insufficient for explaining amounts of flux from these gases.

Denitrifier community composition and diversity was assessed in three soil
treatments varying in average annual N20 flux and history of agriculture. Comparisons
of nirK sequence libraries indicated that community composition of successional sites
abandoned from agriculture 15 years ago still showed impacts from agriculture.
Productivity did not have a significant relationship with measures of denitrifier diversity,
but there was a weak, positive relationship between relative disturbance and diversity. In
addition, a significant, positive relationship between ribosomal RNA (rm) operon copy
number and growth rate of dissirnilatory nitrate reducers was found, indicating that
denitrifiers with low and high rm operon copy numbers differ in ecological strategies.
There was a significant effect of soil treatment on evenness, and there are nirK OTU’s in
the never tilled and historically tilled treatments that may play an important role in
minimizing the amount of N20 flux from sites currently not used for agriculture.

Since members of the order Planctomycetales may play important roles in
nitrogen cycling, a phylogenetic survey of 168 rRNA gene sequences was undertaken to
assess soil planctomycete diversity in soils differing in history of agriculture. Sequences
clustered with the four recognized genera of planctomycetes as well as in clusters outside
those genera. Three sequences were similar to 16S rDNA sequences from organisms
capable of anaerobic ammonia oxidation. Comparative analyses indicated that two soil
treatments that have been used for agriculture are more similar in community
composition and diversity to each other than to a treatment never used for agriculture.
However, there were no significant differences between the three communities,

suggesting that Planctomycetales are marginally affected by agricultural practices.

ACKNOWLEDGEMENTS

When I first started graduate school, my only experience with microbiology was
through the job I had held previously in a lab at a pharmaceutical packaging company. It
was intimidating at first to jump into research in an area I was unfamiliar with, and I owe
a debt of gratitude to all those who helped me initially. From the Schmeznak lab, Joseph
Graber, Joel Klappenbach, Bradley Stevenson, Joel Hashimoto, and Daniel Buckley
showed me the ropes of life in a research lab. Dan especially was incredibly patient
during my rotation in the lab and was the first to teach me about using molecular
techniques. Later members of the lab were also an incredible help. Brendan Keough
worked with me on my BAC library project and as well as being a great technician was a
great friend and gave me a lot of moral support through a tough time during my graduate
experience. I would also like to thank Jorge Rodriguez who was also a great help on the
BAC library project. Stephanie Eichorst was always willing to lend me a helping hand
with experiments and kept lab fun.

My work at the KBS LTER would not have been possible without the help of
many of the graduate students, professors and staff there. In particular, Andrew Corbin
was always accommodating when it came to requests for access to the site and
equipment. Joe Simmons and Greg Parker were very helpful when it came to showing
me how to use equipment and planning the field work involved in the disturbance
experiments discussed in Chapter 2. Stuart Grandy was a wealth of information about
soil and plant processes and helped me understand the “non-microbe” side of the KBS

LTER.

iv

Many of the professors at Michigan State University were crucial to work I
completed here. My two advisors, Tom Schmidt and John Breznak, besides giving great
research advice, have led by example. Each of them is enthusiastic not only about
science and making discoveries, but also about teaching and sharing what they have
learned. The other members of my committee, Terry Marsh, Mike Klug, and Brian
Maurer also provided invaluable advice and insight into my project. John Hoehn was a
collaborator on the modeling study discussed in Chapter 3 which could not have been
completed without his work and helpful discussions.

Lastly, I would like to thank my friends and family. I would not have made it
through graduate school without their support. Julie Hotopp Dunning, Jennifer Gray,
Stephanie Eichorst, and Kristi Whitehead provided me with a social life and fellow grad
students to commiserate with. John Wertz and I shared many a lunch talking about
graduating over Pad Thai. Robin Sutka was a friend and mentor who was extremely
helpful when my project started to change more towards nitrogen cycling in soils. My
mom and sister have been a huge source of support in all my endeavors. Also, my cats
Charcoal, Oliver and Spartacus made excellent lap warmers while writing at the
computer. Finally I would like to thank Matt Chval, who I met at MSU, for his love and
support - I know I was not the most fun to be around while writing.

Thank you everyone!

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... viii
LIST OF FIGURES ........................................................................................................... x
CHAPTER 1: MICROBIAL ECOLOGY OF ORGANISMS CONTRIBUTING TO
LOSS OF NITROGEN FROM SOILS ............................................................................ 1
INTRODUCTION ................................................................................................... l
KBS LTER STUDIES ON MICROORGANISMS INVOLVED IN LOSS OF
NITROGEN FROM SOIL ....................................................................................... 3
SUMMARY ........................................................................................................... 12
THESIS OVERVIEW ............................................................................................ 12
REFERENCES ...................................................................................................... 15

CHAPTER 2: ASSESSING THE INFLUENCE OF CROPS AND
ENVIRONMENTAL FACTORS ON THE FLUX OF GREENHOUSE GASES IN

AGROECOSYSTEMS .................................................................................................... 18
ABSTRACT ................................................................................. 18
INTRODUCTION ......................................................................... 19
MATERIALS & METHODS ............................................................ 21
RESULTS ................................................................................... 26
DISCUSSION .............................................................................. 34
CONCLUSIONS ........................................................................... 41
REFERENCES ............................................................................. 42

CHAPTER 3: COMPOSITION, DIVERSITY, AND NzO FLUX OF

DENITRIFIERS IN AN AGROECOSYSTEM ............................................. 47
ABSTRACT ................................................................................. 47
INTRODUCTION .......................................................................... 48
MATERIALS & METHODS ................................................................................ 51
RESULTS .............................................................................................................. 62
DISCUSSION ........................................................................................................ 81
REFERENCES ...................................................................................................... 92

CHAPTER 4: THE DIVERSITY OF PIANCT 0M YCE TALES IN SOILS

DIFFERING IN HISTORY OF AGRICULTURE ....................................................... 98
ABSTRACT ........................................................................................................... 98
INTRODUCTION ................................................................................................. 98
MATERIALS AND METHODS ......................................................................... 101
RESULTS ............................................................................................................ 104
DISCUSSION ...................................................................................................... l 12
CONCLUSION .................................................................................................... 1 15
REFERENCES .................................................................................................... 1 l7

CHAPTER 5: CONCLUSIONS AND FUTURE DIRECTIONS ............................. 123

SIGNIFICANCE .................................................................................................. 123
SUMMARY ......................................................................................................... 124
DIRECTIONS FOR FURTHER STUDY ........................................................... 127
REFERENCES .................................................................................................... 133
APPENDIX: A STRATEGY FOR CAPTURING PHYLOGENETICALLY
USEFUL INFORMATION IN BAC OR FOSMID LIBRARIES ............................. 137
INTRODUCTION ............................................................................................... 137
DESCRIPTION OF BAC LIBRARY STRATEGY ............................................ 138
CONSTRUCTION OF BAC VECTOR .............................................................. 144
PURE CULTURE AND ENVIRONMENTAL DNA EXTRACTION .............. 145
OPTIMIZATION OF RESTRICTION, LIGATION AND
TRANSFORMATION ........................................................................................ 146
SUMMARY ......................................................................................................... 149
REFERENCES .................................................................................................... 151

vii

LIST OF TABLES
Table 1.1: Treatments used for Microbial Studies at the Kellogg Biological Station
Long-Term Ecological Research Site
Table 2.1: Methods Used or Location of Data Analyzed with Statistical Models
Table 2.2: Multiple Regression Analysis of Factors Influencing C02 Flux
Table 2.3: Multiple Regression Analysis of Factors Influencing CH4 Flux
Table 2.4: Multiple Regression Analysis of Factors Influencing N20 Flux

Table 2.5: Studies Using Multiple Regression Models to Determine Factors Affecting
Greenhouse Gas Fluxes

Table 3.1: Specific Disturbances and Presence (+) or Absence (-) at KBS LTER
Sampling Sites

Table 3.2: pH, Percent Soil Moisture, and N20 Flux of Sampling Sites
Table 3.3: Percent Carbon, Percent Nitrogen and C:N of Sampling Sites
Table 3.4: P-values from I—LIBSHUFF Analysis of Replicate nirK Libraries

Table 3.5: Diversity Statistics and Coverage of nirK Libraries at a 94% Nucleotide
Similarity Cutoff

Table 3.6: Diversity Statistics and Coverage of nirK Libraries at a 97% Nucleotide
Similarity Cutoff

Table 3.7: Diversity Statistics and Coverage of nirK Libraries at a 99% Nucleotide
Similarity Cutoff

Table 3.8: Dominant nirK OTU’s fi'om NT 0-7 cm and their Percent Abundance in HT
and CT at 0—7 cm

Table 3.9: Spearman’s Rank Correlation Coefficient for Diversity Measures (December
2004 nirK Libraries)

Table 3.10: Organisms used to determine the relationship between rm operon copy
number and growth rate

Table 4.1: Diversity statistics for Planctomycetales calculated from 168 rDNA libraries
at a 97% sequence similarity cutoff

viii

Table 5.1: Environmental Surveys for nirK and nirS

ix

LIST OF FIGURES

Figure 1.1: The soil nitrogen cycle. Solid arrows indicate microbial transformations and
dashed lines indicate abiotic processes. DNRA = dissimilatory nitrate reduction to
ammonia.

Figure 1.2: Nitrogen cycling pathways studied at the KBS LTER. Enzymes responsible
for each step in the pathways are named as follows: A) Arno = ammonia
monooxygenase, Hao = hydroxylamine oxidoreductase, and Nor = nitrite oxidoreductase.
B) Nar and Nap = nitrate reductase, Nir = nitrite reductase (specific to denitrification),
Nor = nitric oxide reductase, and Nos = nitrous oxide reductase.

Fig. 2.1: Monthly average greenhouse gas fluxes and associated global warming
potential (GWP) for four treatments at the KBS LTER (from data collected fiom 1992 -
2002). CT (I), HT (0), NT (V ), and DF (0). A) Nitrous oxide flux, B) Methane flux,
and C) Carbon dioxide flux. Error bars represent standard errors.

Figure 3.1: Agarose gel of PCRs of genomic DNA from nirS-bearing Pseudomonas
stutzeri JM300 mixed with environmental DNA from either the NT or CT treatments.
Picogram and nanograrn amounts of DNA listed above lanes refer to the amount of P.
stutzeri genomic DNA present in a PCR containing a total of 50 ng DNA.

Figure 3.2: A visual representation of the similarities between libraries determined by
I-LIBSHUFF analysis. Large blocks represent the two replicate libraries from each
agricultural treatment and soil depth which were not significantly different. Small blocks
represent the libraries fi'om replicate treatments that were significantly different.
Comparisons in I-LIBSHUFF between two libraries involve two library comparisons.
Therefore, depending on the number of libraries represented by each block in the figure,
comparisons between treatments and depths can involve a total of 4 or 8 I-LIBSHUF F
comparisons. Two headed arrows between blocks indicate that two libraries were not
significantly different and a one-headed arrow that a library from one treatment is a
subset of a library in the treatment that is pointed to. No arrows between blocks indicate
that libraries were significantly different from all others.

Figure 3.3: Dendrogram based on a Jaccard dissimilarity matrix at a 94% nucleotide
similarity cutoff. Dendrograrns from the 97% and 99% nucleotide similarity cutoff are
not shown as the same clusters were formed. Scale represents percent dissimilarity.

Figure 3.4: Results of a 2-way ANOVA using data from a 94% nucleotide similarity
cutoff. Significant effects due to agricultural treatment were found for both richness and
evenness at a = 0.05. Graphs show averaged data from 2 replicate libraries from the
same treatment and depth. A) Richness assessed as the number of OTU’s ratified to the
number of clones in the smallest nirK library. B) Simpson’s Evenness. Error bars are
standard errors.

Figure 3.5: Rank/abundance curves of nirK libraries at a 94% nucleotide similarity
cutoff. Treatment replicates that were not significantly different in a I—LIBSHUFF
analysis were combined to obtain an average rank/abundance curve. Treatment replicates
that were significantly different were kept separate. A) Libraries from a depth of 0-7 cm.
B) Libraries from a depth of 13-20 cm.

Figure 3.6: Relationship between Simpson’s Diversity Index and relative disturbance at
94%, 97% and 99% nucleotide similarity cutoffs. Data is fiom libraries created from soil
collected in December 2004.

Figure 3.7: Relationship between dissimilatory nitrate reducer growth rate and rrn copy
number under anaerobic conditions.

Figure 3.8: A) Model describing denitrifier community shifts in response to changes in
an agricultural intensity gradient. The dashed line indicates a speculated community shift
from an HT-like community to a NT-like community given enough time. B) The
community characteristics that were noted in this study and also speculated upon. Points
proven in this study are in black; speculated properties of the denitrifier community are in

grey.

Figure 4.1: Neighbor joining phylogenetic tree constructed from full and partial l6S
rDNA sequences. Numbers in parentheses refer to the number of KBS LTER clones
within each group. Numbers separated by backslashes indicate the number of clones in
each group from a specific soil treatment library (conventional agriculture/abandoned
from agriculture/never tilled). Clones in bold are those clustering most closely with
planctomycetes capable of anammox. The scale bar represents a 10% difference between
nucleotide sequences.

Figure 4.2: Results of I-LIBSHUFF analysis for Planctomycetales 168 rDNA sequence
libraries from three KBS LTER soil treatments (libraries are represented by blocks).
Comparisons between two individual libraries with I—LIBSHUFF actually entail two
statistical comparisons: library X is compared to library Y and Y to X. As such, arrows
point to the library that is being compared to the library being pointed from. Numbers
are p-valucs from comparisons between libraries, and an asterisk (‘) designates p-values
that are significant at a = 0.05.

Figure 4.3: Dendrogram based on a Jaccard dissimilarity matrix calculated at a 97%
nucleotide similarity cutoff. Scale represents percent dissimilarity.

Figure 4.4: Chaol estimates of Planctomycetales richness in three KBS LTER soil
treatments at a 97% 16S rDNA sequence similarity cutoff. (I) Conventional agricultural
treatment, (A) treatment abandoned from agriculture, and (0) never tilled treatment.
Error bars represent 95% confidence intervals.

Figure 4.5: Rank/abundance curves of Planctomycetales 0TU’s defined at a 97%
sequence similarity cutoff. (I) Conventional agricultural treatment, (A) treatment

xi

abandoned from agriculture, and (0) never tilled treatment. Error bars represent 95%
confidence intervals.

Figure 1: Typical orientation of an rRNA operon. Due to the nonpalindromic nature of
the I-CeuI cut site and the orientation of the site in pSuperPhyloBAC, approximately two-
thirds of the 23S rRNA gene and the complete 168 rRNA gene (arrow) of insert DNA
will be ligated into the vector.

Figure 2: Scheme followed during cloning of Xanthomonas campestris pv. campestris
ATCC 33913 DNA into SuperPhyloFOS. A (?) denotes a DNA end resulting from
shearing; therefore it is unknown whether the end has an overhang or is blunt.

Figure 3: Map of the vector pSuperPhyloF OS (not drawn to scale). Section in light gray
is the insert from pSCAN S. Black section is the pCClFOS backbone.

Figure 4: PFG of indirectly extracted soil DNA. Lanes 1 & 17 are Lambda Ladder PFG
Markers; 2 & 16 are Mid-Range H PFG Markers, and 3 & 15 are Low-Range PFG
Markers (all from New England Biolabs). Lanes 4 - 5 are from two extracts of deciduous
forest soil; Lanes 6 - 7 are from two extracts of mid-successional, never-tilled soil. Lanes
8 - 11 are from four extracts of soil abandoned from agriculture; and Lanes 12 - 14 are
from three extracts fi'om conventional agricultural soil.

Figure 5: The dependence of transformation efficiency on voltage and desalting of

sample prior to transformation. (I) 12 kb plasmid control, (V) 79 kb BAC, (o) 115 kb
BAC, (V) desalted 79 kb BAC, and (o) desalted 115 kb BAC.

xii

CHAPTER 1

MICROBIAL ECOLOGY OF ORGANISMS CONTRIBUTING TO LOSS OF
NITROGEN FROM SOILS

Introduction:

Since the Kellogg Biological Station Long-Term Ecological Research (KBS
LTER) site was first set up in 1989, there have been many studies involving soil
microorganisms. The majority have sought to determine what impact Midwestern
agricultural practices have had on bacterial communities, with the ultimate goal of
linking communities with their function. Studies have included those investigating the
effects of isolated or combined agricultural practices on either specific microbial groups
or the microbial community as a whole. The majority of studies concerning specific
groups have focused on microorganisms involved in nitrogen cycling, specifically those
that contribute to the loss of nitrogen from soil ecosystems.

The KBS LTER is composed of eleven treatments that vary in type and degree of
agricultural management as well as successional status (Table 1.1). Some treatments also
contain microplots in which variations on the normally prescribed additions can be
investigated. Treatments are replicated between three and six times and there is a
publicly available database of KBS LTER environmental measurements going back to
1989. This combination of replication and well-documented historical data for the site
has made KBS an ideal setting for studies involving microbes’ role in nitrogen cycling.

In this chapter, I summarize work previously completed at the KBS LTER that
has implications for the work presented in my thesis, discuss microbial pathways

involved in nitrogen cycling, and present an overview of following chapters.

 

fie 28» 8.3 see Bosaaoc ab .6

 

Acme who.» 8.3. 9.315% 88m 859893 688 356883-83 mm mm

ABE—scram .8.“ coho—o no So 63$ $28 gcsvmooa mm “a

Abba—scan com com: 855 32a 333088 357552 HZ mu.

8mg 5 28—50% Sea uncovfinav mac—Q 3223326 E25 bfiotofim Hm E.

£31.. goscucoo mates 8.

$80.» We bog 333.85 much .3qu mOm a.

3:9: 338vo c: 55 5588 33:5 «mogreaoncncmtficov cameo Sufi 33 0M0 E.

€388 32:8 .8§.§3>8-Eo8 :55 33 5 2.

eases Essa seeieaoaoaéoa aroz 4.502 D

9388.. 1353 80:3-§%8-Eoov Bazaar—we 3:369:80 HO 3.
game—omen $35. mat. E scrap—mama

e8: aeoesaaoc mm: mam

an eoaoaom aoaoaoom confines eoeam aoaoaa $23 as a seam 33822 as eon: aeoaaae ”Z 23.

KBS LTER Studies on Microorganisms Involved in Loss of Nitrogen from Soil:

The majority of microbiological studies at the KBS LTER have involved those
organisms that participate in the nitrifying and denitrifying portion of the nitrogen cycle
(Fig. 1.1). These two groups are particularly important because their combined activity

can be responsible for losses of nitrogen from agricultural soil as high as 60% [1].

Nimfiers: Nitrification is performed by two different groups of bacteria, those
capable of ammonia oxidation to nitrite and those that oxidize nitrite to nitrate (Fig.
1.2A). Until recently, it was thought that only certain members of the y- and B-
proteobacteria were capable of nitrification. However, members of the kingdom
Chrenarchaeota also are capable of ammonia oxidation [2]. Expression of
Chrenarchaeota amoA-like genes was found to occur in soil [3], indicating that this group
may also be important in nitrogen cycling [4].

Nitrite and nitrate are highly mobile forms of nitrogen due to their negative
charge, and will leach out of agricultural soil and into water resources. Nitrite can be
mutagenic when it forms nitrous acid in the environment and small amounts have
antimicrobial effects in acidic soils [5]. Nitrification also causes direct losses of nitrogen
from soil by production of the greenhouse gas N20 and indirect losses by creating
nitrogen species subsequently used in denitrification. Of the nitrifiers, ammonia
oxidizing bacteria (AOB) have been the focus of KBS studies.

The rate of nitrification is affected by ammonium concentration, temperature,
moisture, and oxygen concentration [6, 7]. While many studies have focused on

nitrification rates, there are few that focus on how nitrifier diversity and community

 

NO, N20, N2

A

 

¥ /
[“2 xk / ”Twat?”

/ \

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Nitrogen \
Fixatlon I’ \
, I \
* Nitrificationl \i
Organic 7 ----- * NH4+-—-> N02. —> N03- '—> N02-

M tt - - - . . . 7
R-aNl-el; I Mineralization] f‘ I Denltrlficatlon I
—— N02“

| DNRA]

 

Fig. 1.1: The soil nitrogen cycle. Solid arrows indicate microbial transformations and
dashed lines indicate abiotic processes. DNRA = dissimilatory nitrate reduction to

ammonia.

A. Nitrification

NH3 '—> NHon —* N02- —'* N03-
Arno Hao Nor

B. Denitrification

NOg' -—> N02' —> NO —* N20 —> N;
Nar Nir Nor Nos
Nap

Fig 1.2: Nitrogen cycling pathways studied at the KBS LTER. Enzymes responsible for
each step in the pathways are named as follows: A) Amo = ammonia monooxygenase,
Hao = hydroxylamine oxidoreductase, and Nor = nitrite oxidoreductase. B) Nar and Nap
= nitrate reductase, Nir = nitrite reductase (specific to denitrification), Nor = nitric oxide
reductase, and Nos = nitrous oxide reductase.

structure affect those rates or how agricultural treatments affect community structure.
Therefore the roles of fertilization and tillage as well as seasonal effects on the nitrifier
community have been investigated at KBS since these are the primary factors in
agricultural systems that may influence nitrifiers.

Fertilization has an effect on the abundance of nitrifiers. Initial most probable
number (MPN) experiments with media containing different concentrations of
ammonium sulfate indicated that AOB communities in agricultural fields subjected to
regular ammonia-containing, nitrogenous fertilizer applications are more tolerant of high
concentrations of ammonium and are higher in number compared to communities present
in soil that do not receive regular nitrogen inputs [8]. Later studies sought to confirm
these results using both MPN counts and competitive PCR (cPCR) [9, 10]. MPN
underestimated AOB numbers as compared to cPCR, and in these studies no difference
between population numbers was found with MPN. However with cPCR, fertilization
again increased nitrifier numbers. Despite the increase, no correlation between potential
nitrification rate and ACE numbers was found. This was also the case in a study on non-
KBS soils in which AOB community size and nitrification were monitored before and
after fertilizer addition. Fertilization caused an increase in nitrification rate within 3 days
of fertilization, but the ACE community size did not increase significantly until 6 weeks
had passed and nitrification rates had decreased [1 1]. This indicates that the members of
the ACE community or physical factors in the field are more important in determining
nitrification rates than the sheer number of organisms. Indeed, a study conducted using
plots from the KBS LTER and Living Field Laboratory (LF L) aimed specifically at

investigating physical factors influencing nitrification rates found a seasonal pattern to

nitrification potentials, but that the cropping system had no effect [12]. This suggests that
seasonal effects, such as temperature or soil moisture, are important drivers of
nitrification rates.

The results of KBS studies on the effect of agriculture on nitrifier diversity and
community composition indicate that there is a long-lasting effect of agricultural use.
Bruns and colleagues [8] found A08 in a conventionally treated agricultural field were
less diverse than those from a never-tilled field based on DGGE and sequencing of A03
168 rRNA gene clones. Sequences from Nitrosospira cluster 3 were the only ones found
in the agricultural field while sequences from an additional three AOB groups were found
in the never-tilled field. Phillips and colleagues [10] found no detectable differences in
AOB communities from the same agricultural fields or a site abandoned from agriculture
in 1989, using the same techniques. Sequences from Nitrosospira cluster 3 dominated at
both sites in this study. The treatment abandoned from agriculture was studied in 1994
and 1995; between 5 and 6 years after abandonment. The fact that the A03 community
after this amount of time was still no different from that in the agricultural treatment
suggests that the community had not recovered from the impact of agricultural practices.

Later studies from outside the KBS LTER supported these findings. Fertilization
causes changes in AOB community composition, but these changes occur at a slow rate,
most likely due to the inherently slow growth of the microorganisms [13-15]. After
fertilizer was added to soil microcosms, significant changes in the DGGE patterns of the
amoA gene were not detectable until 16 to 20 weeks of incubation had occurred [14], and
in an earlier study by the same group there was no difference between communities in

soils that received fertilizer and did not receive fertilizer after 4 weeks [13]. Changes in

microcosm communities receiving a concentrated pulse of ammonium would be expected
to occur more quickly than under field conditions. Therefore, it is possible that 5 to 6
years after abandonment from agriculture was not enough time to see a shift in KBS
nitrifier communities in the field.

In summary, there is a trend of increasing AOB numbers with fertilization, and
fertilization selects for AOB that are tolerant to high concentrations of ammonium. Also,
nitrification potentials are higher in agricultural treatments, and this cannot be explained
solely by the difference in nitrifier numbers. This may be due to the combination of
physical factors differing between fields and differences in the diversity and composition
of the nitrifier community. In fact, in a recent study physiological differences between
members of Nitrosospira cluster 3 accounted for differences in the time needed to initiate
nitrification [16]. In natural soils, certain members of the group were sensitive to high
ammonium concentrations, and others were tolerant, causingla delay in nitrification until
tolerant organisms could grow to sufficient population size. KBS LTER studies also
demonstrated that agriculture can induce shifts in AOB community structure and that

these changes are long-lasting.

Denimfiers: Denitrification is a form of dissimilatory nitrate reduction performed
by a phylogenetically diverse group of microbes that includes Proteobacteria, Gram
positive bacteria, Archaea, and fungi. Denitrifiers are typically heterotrophic,
facultatively anaerobic organisms that are capable of using nitrate as an electron acceptor
under anaerobic conditions. During denitrification, nitrate is reduced in stepwise manner

to nitrite, nitric oxide, nitrous oxide and dinitrogen gas (Fig. 1.2B). This process serves

as a source and sink of nitrous oxide (N20), a potent greenhouse gas. As well as
contributing to gas emissions, deniuification removes nitrogen from soil by converting it
to a readily diffusible gaseous product. An average of 20-30% of nitrogenous fertilizer
added to an agricultural field will be lost to denitrification [17]. Nitrifiers also carry out
denitrification under low oxygen concentrations [18-21]. Studies at the KBS LTER have
not focused on nitrifier denitrification and there is some question as to its importance,
since measurements of N20 production by nitrifier denitrification range from
insignificant amounts [22] to 30% of the total N20 produced [23].

As with nitrifiers, studies on denitrifiers at the KBS LTER have sought to link
community composition and diversity with ecosystem processes, most notably, that of
N20 flux. Denitrification rates as well as the proportion of denitrification resulting in
production of N20, or N20 mole fraction, are influenced by pH, soil moisture, C/N ratio,
oxygen concentration, and temperature [7, 24, 25]. KBS LTER studies have supported
the hypothesis that in addition to physical factors, denitrifier community composition
influences N20 production.

Experiments at the KBS LTER investigating the response of denitrifiers to the
environmental factors of pH, 02 concentration, and moisture history found differences
that help to explain differences seen in N20 flux between a conventional agriculture
treatment, a treatment abandoned from agriculture since 1989, and a never-tilled
treatment. The site subjected to conventional agricultural practices has an annual average
N20 flux about three times higher than the two sites not currently used for agriculture
[26]. The agricultural treatment was found to have denitrifiers with N20 production

enzymes (Nar, Nir and Nor) that were more sensitive to 02 concentration than those of

the never-tilled treatment. Conversely, the never-tilled treatment community had N20
production enzymes that were more sensitive to pH as compared to the agricultural
treatment. In both communities, Nos, which destroys N20, was sensitive to 02 but not
pH. Nos from the agricultural treatment was significantly more sensitive to 02 than that
from the never-tilled treatment. To follow up this study, denitrifiers were cultivated from
both soil treatments, and the sensitivity of Nos from each isolate to 02 was tested [27].
Nos of the isolates had a wide range of sensitivities to 02, but there was no difference
between activities at the community level. However, this survey of denitrifiers was
limited by the fact that only cultivated denitrifiers were examined. In a later comparison
of the N20 mole fraction ([N20]/[N20 + N2]) produced by the agricultural treatment and
the treatment abandoned from agriculture, there was a significant difference in N20 mole
fraction, but not in total denitrification [25]. There was also a strong influence of soil
moisture history. The agricultural treatment had a significantly higher N20 mole fraction
with no pre-wetting. Pre-wetted soil from both treatments did not differ in N20 mole
fi'action, suggesting that the denitrifiers in the site not currently used for agriculture had
Nos that was able to persist for a longer time under dry conditions than that from the
agricultural treatment.

The diversity and composition of denitrifier communities at the KBS LTER have
also been assessed and used to investigate further the link between diversity and
ecosystem function as well as factors that impact community structure. Treatments
investigated included some of the same used to test differences in response to
environmental factors; the conventional agriculture treatment and plots that have never

been tilled or used for agriculture.

10

There are distinct differences between the denitrifier communities of the
agricultural and successional treatments. Two studies focusing solely on comparing the
agricultural and never-tilled plots found that the never-tilled plot denitrifiers were less
diverse than those of the agricultural plot and that their composition differed [27, 28].
These studies employed both cultivation and molecular methods. Of cultivated
denitrifiers, a total of 27 taxa were found, but only 12 were shared between treatments.
The number of cultivated isolates fi'om each treatment indicated that there were a higher
number of denitrifiers in the agricultural treatment. Organisms from the agricultural
treatment tended to be members of the a- and B-proteobacteria while organisms fi'om the
never-tilled treatment were mostly y-proteobacteria. The diversity of isolates was higher
in the agricultural treatment as compared to those from the never-tilled treatment [27].

A molecular survey of the same treatments confirmed these results. Restriction
Fragment Length Polymorphism (RFLP) of 710.92 was used to assess diversity, and 182
distinct RFLP patterns were identified. As in the previous study, few groups (only 8
RF LP patterns) were shared between communities and the diversity of the agricultural
treatment was higher than that of the never-tilled. This was due to the richness and
evenness of the agricultural field being higher. Evenness is a measure of the degree a
community is dominated by particular species. The higher evenness of the agricultural
treatment indicated that few RFLP patterns dominated the population. Indeed, the group
with the highest fi'equency made up less than 5% of the total while the never-tilled field
had one group that made up 32% of the RF LP patterns found [28].

The KBS LTER studies on links between the denitrifier community and

ecosystem function were among the first to show that communities can differ in their

11

physiological potentials and response to their environment, which in turn affects
frmctions such as N20 flux. Other studies on denitrifiers have confirmed these results
[29-32]. For instance, not only are there differences in denitrifier communities between
soil treatments, but there are also differences within treatments that contribute to
differences in ecosystem function. An example of this are the differences noted by
Cheneby and colleagues [32] between communities in maize rhizosphere and non-
rhizosphere soil. Denitrifiers isolated from the rhizosphere were less diverse than those
from the surrounding soil, and the dominant rhizosphere denitrifiers were not able to

reduce N20 to N2, whereas bulk soil isolates produced N; as their denitrification product.

Summary:

KBS LTER studies on nitrifying and denitrifying bacterial communities have
demonstrated that they will respond to changes in physical factors caused by agricultural
practices, causing them to differ from communities in non-agricultural soils in
composition, diversity, and abundance. In turn, microbial communities can differ in their
physiological potential and response to environmental conditions. This ultimately results
in differences in various ecosystem functions. Published work suggests that microbial
communities as well as environmental conditions should be taken into account when

investigating ecosystem functions such as gas flux or nutrient cycling.

Thesis Overview:

Nitrous oxide flux from soil tends to be spatially and temporally variable, making

predictions of total flux difficult. Therefore, in Chapter 2, a statistical modeling approach

12

was taken to determine how much of the variability in flux of N20 and of two other
greenhouse gases (CO2 and CH4) could be explained from four treatments varying in
history of agriculture at the KBS LTER. The amount of explained variation was
generally low, with CO2 flux having the greatest amount of explainable variation. It was
concluded that the high amount of explainable variability in CO2 flux was due to the fact
that this gas is a byproduct of all types of heterotrophic metabolism in soil, making its
production responsive to any factors affecting microbial metabolism in general. Nitrous
oxide and CH.:, conversely, are produced or consumed by microorganisms with
specialized metabolisms, causing environmental variables alone to be insufficient to
explain amounts of flux from these gases.

The findings presented in Chapter 2 suggest that differences in the composition
and diversity of microorganisms responsible for greenhouse gas fluxes are important for
understanding variation in this ecosystem function. Therefore, in Chapter 3, denitrifier
community composition and diversity was assessed in three soil treatments that varied in
their average annual N2O flux and history of agriculture. Differences in community
composition and diversity were found; therefore the effect of disturbance and
productivity as well as a possible ecological strategy of these organisms was investigated.

In Chapter 4, the community composition and diversity of microorganisms
belonging to the order Planctomycetales was determined. There have been numerous
studies on planctomycetes from aquatic habitats, and recently isolates from soil have
been cultivated. Despite this, little is known about the ecological role of this diverse
group. One intriguing recent development was the discovery that organisms capable of

anaerobic oxidation of ammonia (anammox) belong to the Planctomycetales [33]. If

13

anammox were occurring in soil, it would have implications for the nitrogen cycle in soil
as the end product of this reaction is dinitrogen gas; making this another microbial
pathway resulting in loss of nitrogen from soils.

In Chapter 5, the conclusions from this thesis are presented along with proposed
experiments that would continue to further our knowledge regarding the ecology of

microorganisms involved in the soil nitrogen cycle.

14

REFERENCES:

l.

10.

ll.

Robertson, G.P., Nitrogen use efliciency in row-crop agriculture: crop nitrogen
use and soil nitrogen loss, in Ecology in Agriculture, L.E. Jackson, Editor. 1997,
Academic Press: San Diego. p. 347-365.

Konneke, M., A.E. Bernhard, J .R. de la Torre, C.B. Walker, J .B. Waterbury, and
DA. Stahl, Isolation of an autotrophic ammonia-oxidizing marine archaeon.
Nature, 2005. 437(7058): p. 543-546.

Treusch, A.H., S. Leininger, A. Kletzin, S.C. Schuster, H.P. Klenk, and C.
Schleper, Novel genes for nitrite reductase and Amo-related proteins indicate a

role of uncultivated mesophilic crenarchaeota in nitrogen cycling. Environmental
Microbiology, 2005. 7(12): p. 1985-1995.

Nicol, G.W. and C. Schleper, Ammonia-oxidising Crenarchaeota: important
players in the nitrogen cycle? Trends in Microbiology, 2006. 14(5): p. 207-212.

Bancroft, K., I.F. Grant, and M. Alexander, Toxicity of N02: efi'ect of nitrite on
microbial activity in an acid soil. Applied and Environmental Microbiology,
1979. 38(5): p. 940-944.

Robertson, G.P., Factors regulating nitrification in primary and secondary
succession. Ecology, 1982. 63(5): p. 1561-1573.

Granli, T. and QC. Bockrnan, Nitrous oxide from agriculture. Norwegian Journal
of Agricultural Sciences Supplement, 1994. 12: p. 7-128.

Bruns, M.A., J .R. Stephen, G.A. Kowalchuk, J.I. Prosser, and EA. Paul,
Comparative diversity of ammonia oxidizer 16S rRNA gene sequences in native,
tilled, and successional soils. Applied and Environmental Microbiology, 1999.
65(7): p. 2994-3000.

Phillips, C.J., E.A. Paul, and J .I. Prosser, Quantitative analysis of ammonia
oxidising bacteria using competitive PCR. FEMS Microbiology Ecology, 2000.
32: p. 167-175.

Phillips, C.J., D. Harris, S.L. Dollhopf, K.L. Gross, 1.1. Prosser, and EA. Paul,
Efi'ects of agronomic treatments on structure and fitnction of ammonia-oxidizing
communities. Applied and Environmental Microbiology, 2000. 66(12): p. 5410-
5418.

Mendum, T.A., R.E. Sockett, and PR. Hirsch, Use of molecular and isotopic
techniques to monitor the response of autotrophic ammonia-oxidizing populations

of the ,8 subdivision of the Class Proteobacteria in arable soils to nitrogen
fertilizer. Applied and Environmental Microbiology, 1999. 65(9): p. 4155-4162.

15

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Fortuna, A., R.R. Harwood, G.P. Robertson, J .W. Fisk, and EA. Paul, Seasonal
changes in nitrification potential associated with application of N fertilizer and

compost in maize systems of southwest Michigan. Agriculture Ecosystems &
Environment, 2003. 97(1-3): p. 285-293.

Avrahami, S., R. Conrad, and G. Braker, Efl'ect of soil ammonium concentration
on N20 release and on the community structure of ammonia oxidizers and
denitrifiers. Applied and Environmental Microbiology, 2002. 68(11): p. 5685-
5692.

Avrahami, S., W. Liesack, and R. Conrad, Eflects of temperature and fertilizer on
activity and community structure of soil ammonia oxidizers. Environmental
Microbiology, 2003. 5(8): p. 691-705.

Avraharni, S. and R. Conrad, Patterns of community change among ammonia

oxidizers in meadow soils upon long-term incubation at difi’erent temperatures.
Applied and Environmental Microbiology, 2003. 69(10): p. 6152-6164.

Webster, 0., TM. Embley, T.E. Freitag, Z. Smith, and J .I. Prosser, Links between
ammonia oxidizer species composition, functional diversity and nitrification
kinetics in grassland soils. Enviromnental Microbiology, 2005. 7(5): p. 676-684.

Tiedje, J.M., Ecology of denitrrfication and dissimilatory nitrate reduction to
ammonium, in Biology of Anaerobic Organisms, R. Mitchell, Editor. 1988, John
Wiley & Sons: New York. p. 179-244.

Goreau, T.J., W.A. Kaplan, S.C. Wofsy, M.B. McElroy, F.W. Valois, and SW.
Watson, Production of NO2’ and N20 by Nimfling Bacteria at Reduced
Concentrations of Oxygen. Applied and Environmental Microbiology, 1980.
40(3): p. 526-532.

Hynes, R.K. and R. Knowles, Production of Nitrous-Oxide by Nitrosomonas-
Europaea - Effects of Acetylene, Ph, and Oxygen. Canadian Journal of
Microbiology, 1984. 30(11): p. 1397-1404.

Poth, M. and DD. Focht, N—I5 Kinetic-Analysis of N20 Production by
Nitrosomonas-Europaea - an Examination of Nitrifier Denitrtfication. Applied
and Environmental Microbiology, 1985. 49(5): p. 1134-1141.

Wrage, N., G.L. Velthof, M.L. van Beusichem, and O. Oenema, Role of nitrrfier
denitrtfication in the production of nitrous oxide. Soil Biology and Biochemistry,
2001. 33: p. 1723-1732.

Robertson, GP. and J .M. Tiedje, Nitrous-Oxide Sources in Aerobic Soils -

Nitrification, Denitrrfication and Other Biological Processes. Soil Biology &
Biochemistry, 1987. 19(2): p. 187-193.

16

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

Webster, EA. and D.W. Hopkins, Contributions from drflerent microbial
processes to N20 emission from soil under different moisture regimes. Biology
and Fertility of Soils, 1996. 22(4): p. 331-335.

Firestone, MK. and EA. Davidson, Microbiological basis of NO and N20
production and consumption in soil, in Exchange of Trace Gases between
Terrestrial Ecosystems and the Atmosphere, M.O. Andreae and D.S. Schimel,
Editors. 1989, John Wiley & Sons: New York. p. 7-21.

Bergsma, T.T., G.P. Robertson, and NE. Ostrom, Influence of soil moisture and
land use history on denitrification end-products. Journal of Environmental
Quality, 2002. 31(3): p. 711-717.

Robertson, G.P., E.A. Paul, and RR. Harwood, Greenhouse gases in intensive
agriculture: contributions of individual gases to the radiative forcing of the
atmosphere. Science, 2000. 289: p. 1922-1925.

Cavigelli, MA. and G.P. Robertson, Role of denitrtfier diversity in rates of
nitrous oxide consumption in a terrestrial ecosystem. Soil Biology and
Biochemistry, 2001. 33: p. 297-310.

Stres, B., I. Mahne, G. Avgustin, and J .M. Tiedje, Nitrous Oxide Reductase
(nosZ) Gene Fragments Difi’er between Native and Cultivated Michigan Soils.
Applied and Environmental Microbiology, 2004. 70(1): p. 301-309.

Holtan-Hartwig, L., P. Dorsch, and LR. Bakken, Comparison of denimfiing
communities in organic soils: kinetics of NO3' and N20 reduction. Soil Biology &
Biochemistry, 2000. 32(6): p. 833-843.

Holtan-Hartwi g, L., P. Dorsch, and LR. Bakken, Low temperature control of soil
denitrifying communities: kinetics of N 20 production and reduction. Soil Biology
& Biochemistry, 2002. 34(11): p. 1797-1806.

Rich, J.J., R.S. Heichen, PJ. Bottomley, K. Cromack, Jr., and DD. Myrold,
Community Composition and Functioning of Denitrifying Bacteria from Adjacent
Meadow and Forest Soils. Applied and Environmental Microbiology, 2003.
69(10): p. 5974-5982.

Cheneby, D., S. Perrez, C. Devroe, S. Hallet, Y. Couton, F. Bizouard, G. Iuretig,
J .C. German, and L. Philippot, Denitrijying bacteria in bulk and maize-
rhizospheric soil: diversity and N20-reducing abilities. Canadian Journal of
Microbiology, 2004. 50(7): p. 469-474.

Strous, M., J.A. Fuerst, E.H.M. Kramer, S. Logemann, G. Muyzer, K.T. van de

Pas-Schoonen, R. Webb, J .G. Kuenen, and SM. J etten, Missing Lithotroph
Identified as New Planctomycete. Nature, 1999. 400: p. 446-449.

17

CHAPTER 2

ASSESSING THE INFLUENCE OF CROPS AND ENVIRONMENTAL
FACTORS ON THE FLUX OF GREENHOUSE GASES IN AGROECOSYSTEMS

These results were submitted for publication in the article: Huizinga K.M., U.Y. Levine,
J.P. Hoehn, and T.M. Schmidt. 2006. Assessing the Influence of Crops and
Environmental Factors on the Flux of Greenhouse Gases in Agroecosystems.

Biogeochemistry.

ABSTRACT:

In an effort to better understand variation in the production and consumption of
greenhouse gases by microbial communities in soils of agroecosystems, data from four
soil ecosystems at the Kellogg Biological Station Long Term Ecological Research Site
were analyzed. The data spanned 6 to 9 years for soils that varied from a standard crop
rotation of corn, soybeans, and wheat to fallow land and deciduous forest. Gas fluxes
were modeled using a production function approach where gas flux is a function of
current year management, historical management, and environmental factors. Linear
production function parameters were estimated using ordinary least squares with robust
standard errors. Independent variables explained between 8 and 50% of the variation in
gas flux, with carbon dioxide flux consistently having the greatest amount of explainable
variation (29 to 50%). We conclude that since CO2 is a byproduct of all types of
heterotrophic respiration in soil, its flux will be responsive to factors that enhance
microbial metabolism in general, and that this phenomenon explains the high coefficients

of determination for explaining variation in C02 flux at the KBS LTER. Conversely,

18

since the consumption of CH4 and production of N20 require the activity of microbes
with specialized metabolic pathways, environmental parameters alone explain a smaller
percentage of the flux of these gases. There were also significant effects on gas flux
associated with specific crops, with fluxes roughly following the pattern wheat > soybean
> corn. In addition, the treatment with the highest level of variation accounted for by
environmental factors alone was the late-successional, deciduous forest (19 to 50%).
Future modeling efforts will likely be enhanced by including measures of microbial
diversity in soil, particularly those microbes involved in methane consumption and

nitrous oxide production.

INTRODUCTION:

In terrestrial ecosystems, the exchange of carbon dioxide (CO2), methane (CI-I4)
and nitrous oxide (N 20) between the soil and atmosphere has important effects on
ecosystem services such as climate regulation and nutrient cycling. The production and
consumption, or flux, of these greenhouse gases varies seasonally and spatially across
landscapes, and is influenced directly by land use, particularly agriculture. For instance,
the majority of labile organic matter in soil is oxidized to CO2 and lost to the atmosphere
during the first few years alter native lands are converted to agriculture [1, 2]. Managing
land for agriculture also decreases the amount of CH4 that is consumed by soil
microorganisms [3], and increases the emission of N20 from soil [4, 5]. Identifying
environmental factors that influence the exchange of these greenhouse gases between
soils and the atmosphere would enhance predictive models of gas flux, which are

important in evaluating ecosystem services provided by agriculture. In addition, it may

19

also suggest strategies for mitigating the detrimental impact of greenhouse gases
produced by agriculture.

Inventories of agricultural gas fluxes used in climate change assessments and
policy analysis are typically developed using biogeochemical process models. Process
models such as DAYCENT [6, 7] or DNDC [8], are generally composed of several
submodels for predicting the effects of many interrelated biological and physical
processes on gas flux. These models depend on accurate physical measurements of
environmental variables to calculate a predicted amount of gas production or
consumption. While techniques for measuring environmental data have become more
accurate over time, models are still limited in their predictive power due to the many
parameter estimates and assumptions required to construct them. For instance, in terms
of assumptions, IPCC guidelines [9] consider all agricultural systems to be equivalent,
which does not take into account spatial and temporal variability, or differences in the
soil microbial communities that are ultimately responsible for the production and
consumption of greenhouse gases.

In this chapter, a production function approach was used to evaluate the effects of
environmental and crop management variables on the flux of C02, N20, and CH4. An
empirical assessment of the importance of these variables prior to their use in process
models will serve to improve gas flux predictions. Estimation of the production frmction
parameters is based on data covering approximately 9 years at the Kellogg Biological
Station Long Term Ecological Research Site (KBS LTER) in Michigan. The production
function approach posits that gas flux is a function of environmental factors, crop

management values, and unmeasured factors represented by a stochastic error. This

20

approach allows us to assess the impact of the environmental and management factors on
gas flux, both individually and as a group of measured variables. The variation in flux
that is unexplained by the measured variables also provides an upper bound assessment
of the effect of unmeasured variables, such as those describing the microbial community.
Results indicate that the included measured variables explain 50% or less of the variation
in gas flux. The amount of explained variation varies by agroecosystem type and by type
of flux and are greatest for CO2 and least for N20. The results suggest that while
environmental and management factors are helpful in understanding a minor portion of
gas flux, smaller error predictions are only likely to come with a better understanding and
measurement of extant unmeasured variables, such as the composition, diversity, and

structure of soil microbial communities.

MATERIALS & METHODS:
Study Site

Four treatments were the focus of this study: a conventional agriculture site that
receives amounts of fertilizer, herbicide, and conventional tillage (CT) that are typical for
the Midwest region and is on an annual rotation of corn (Zea mays L.) - soybean (Glycine
max L.) - wheat (Triticum aestivum L.); an agriculture site with a history of tillage (HT)
that was abandoned from agricultural use in 1989; a never-tilled site (NT) maintained in a
mid-successional plant community by annual mowing and a late-successional deciduous
forest (DF). The KBS LTER is set up in a randomized block design and data from four
replicates of the CT, HT, and NT treatments was retrieved, as was data from the three DF

replicates. Additional details on the KBS LTER can be found at http://lter.kbs.msu.edu/.

21

The Production Function for Gas Flux

The production fimction approach posits that gas flux is a function of
environmental factors, crop management, and a stochastic term. This allows one to
quantitatively estimate the partial effect of each measured environmental and
management factor on the ecological service of interest, in this case, gas flux. In
addition, the approach indicates the importance of the measured factors in explaining the
variation in greenhouse gases relative to the influence of the unmeasured factors
represented by the stochastic error term.

The estimated production function is specified as a linear model:
(1) g = a + Bx + e
where g is gas flux and x is a K-element vector of measured environmental and

management variables. The quantities a and B are parameters to be estimated and e is
the stochastic term. The a represents the mean value of g when the environmental and

management variables are equal to zero. The i, i = (1,..., K) represents the partial effect

of the ith environmental or management variable on gas flux; it describes how gas flux
changes for a one unit change in x,. a and B are reported as CO2 equivalents, or global
warming potential (GWP), using IPCC conversion factors for a 20-year time span [9].
Conversion to CO2 equivalents allows for comparisons of the contribution of a particular
factor between modeled fluxes. The formulas used to calculate CO2 equivalents were
those reported in the supplemental material of a study by Robertson and colleagues [5].
Each factor (x,) was selected based on its known propensity to influence gas
fluxes from soil. Variables fall into one of three categories: current year management,

historical management, and environmental factors. Current year management factors

22

apply only to the conventional agricultural treatment and include the effect of different
crop types and fertilization (assuming a 30 day residence time). Crops vary in the nature
and amount of organic compounds released through root exudates and in their chemical
composition: both have the potential to influence the structure and frrnction of microbial
communities in soil. Fertilization provides nitrogen not only to plants, but also to
microbes in soil and so has the potential to stimulate microbial metabolism and influence
the production of CO2 and N20. Historical management factors were considered to be
those characteristics of soil that reflect past land use. These factors were percent soil
moisture, water filled pore space (WFPS), nitrate concentration, ammonium
concentration, percent total carbon, and carbon/nitrogen (C/N) ratio. Water filled pore
space was calculated using the bulk density of each soil treatment replicate, percent soil
moisture, and assuming an average soil particle density of 2.65 g-cm'3 [10]. Moisture
affects microbial activity by influencing oxygen concentration and availability of
nutrients. In particular, high soil moisture will result in anoxic conditions that favor
fermentation and denitrification, but not methane oxidation or aerobic respiration.
Nitrogen and carbon will stimulate bacterial growth as mentioned above, and different
microorganisms will be favored as the relative amounts of each differ. The
environmental factors used in the models were cumulative precipitation from three days
before measurement of gas fluxes, cumulative precipitation from seven days before
measurement of gas fluxes, maximum air temperature, and minimum air temperature.
Precipitation will affect soil moisture and temperature will affect microbial activity.
Microbial metabolism increases as temperature rises which increases the rate of gas

production or consumption. Records of CO2, CH4, and N20 fluxes along with

23

management and environmental variables from the KBS LTER were downloaded from
the KBS LTER website (http://lter.kbs.msu.edu /Data/DataCatalog.html). The units for
each variable and method of collection are recorded in Table 2.1.

Ifthe measurement of any factor was not available on the day that gas fluxes were
measured, data from measurements closest to that date were used as follows: if a gas
sampling date fell exactly between two field measurements, they were averaged. In
addition, for the conventional agriculture treatment measures of UN ratio, the time of
fertilization was used to determine which C/N data point to assign to a particular gas flux
measurement. For example, if C/N measurements were taken 5 days before and 2 days
after gas measurements, but fertilization took place one day after, the C/N measurement
from 5 days before the gas measurement was used to avoid falsely biasing the data due to
fertilization.

Data sets for each soil treatment were assembled to maximize the number of data
points to be included in the analyses. As such, data sets did not all contain data from the
exact same days. The conventional agriculture data set was comprised of 408 days of
observations collected from 1992 - 2001 (one outlying N20 flux measurement was
removed), the historical agriculture data set contained data from 351 days between 1992 -
2000, the never tilled treatment consisted of 275 days of data collected from 1992 - 1998,
and the data set for the deciduous forest consisted of 223 days of data collection from
1993 - 2000. Gas measurements at the forest treatment were performed so that two
samples per replicate were taken on each day of sampling. The two flux measurements
were averaged as the corresponding physical data was performed once for each forest

replicate.

24

 

33% 2a

 

mama—00 883v 825x95 23.82.88 5w
538m
5583 $5 may see 8283. as E8 5 Ba 3 83298:. Beogesam
282 20
a: 83588 ca 2: .. £8 .5 mo 3 :85 .x.
:8 be 32 w: comgcooaou §EoEE<
.3 ages
emanate—co 93 558.86 5v— :8 5 w \ Z w: coup—E880 38:2
2 a .888 as; =3 0808380 2: ._ as be scam 3 ages: :8 .x. assumes: nausea
«53 Eugene
ms 3825—9” m9: ma 5 vow—coed ocoz e80 3ch?an 30> ago
280% me 5:33 8 com: ween—z 9:5 88am b.0330 88am

 

2082 285% a? 332 98 a 8283 s was 8502 as as;

25

To determine which factors had significant effects on greenhouse gas fluxes,
models were set up using a linear production function in the Stata program (StataCorp,
College Station, TX). Separate analyses were performed for each treatment and
greenhouse gas. In each analysis, gas flux was the dependent variable (g) and the

previously listed environmental measures were independent variables (xi ). The amount

of variation in gas flux explained by the environmental and management factors relative

to the stochastic term was measured by R2. R2 is equal to 1 - 08/ 08 where 0', is the

variance of c and 0'g is the variance of g.

RESULTS:
Seasonal Greenhouse Gas Fluxes

Of the four treatments, soils of the conventional agriculture treatment had the
highest N20 flux throughout the year, with peak production in April and May (Fig 2.1A).
There was a seasonal peak of flux occurring in May and June in the deciduous forest, but
as previously reported, the annual average flux of N20 from the historical agriculture,
never tilled soil, and deciduous forest treatments was similar [5]. The never tilled and
forest treatments both had the highest rate of CH4 consumption during the summer and
early fall; the conventional agriculture and historical agriculture treatments had the
lowest capacity for CH4 consumption throughout the year with little variation in flux
(Fig. 2.13). The flux of CO2 from the historical agriculture and never tilled treatments
was highest in late spring, and then peaked again in August (Fig. 2.1C). The C02 flux
from both the conventional agriculture and forest treatments was lower overall and also

peaked during the summer.

26

E“

N20-N (g-ha"-d")

 

 

O '5 8 8 3 3 8 3 8
GWP(gC02.m2.y')

 

 

 

 

 

 

 

 

B. 2
.2 A
7A .5 ”3.,
l? -10 ”a
M
43‘- -l4 5v
39 -18 o
L.) 22 e
5' r ' S
-101 "26 o
. ..30
-121 . - - . - “34
Feb Apr Jun Aug Oct Dec
C“ 70: .9
1‘ 60‘ .8 .f‘
3 50: 17 a?
g 1 46 E.
o 40‘ N
e . ‘5 8
U 30. .4 0,,
'~ ' 3 '5
O 20- '
u j .2 S
10 . l O
0 V I— V I V I V I V I 0
Feb Apr Jun Aug Oct Dec

Month

Fig. 2.1: Monthly average greenhouse gas fluxes and associated global warming
potential (GWP) for four treatments at the KBS LTER (from data collected from 1992 -
2002). CT (I), HT (C), NT ( ), and DF (13). A) Nitrous oxide flux, B) Methane flux,
and C) Carbon dioxide flux. Error bars represent standard errors.

27

Current Year Management Factors

As the conventional agriculture site was the only treatment in this study currently
used for agriculture, it is the only one to which current year management factors applied.
Modeling of all gases for this treatment was performed using the corn crop as the
baseline, causing the effect of this crop to be absorbed into the intercept coefficient. Both
soybean and wheat increased CO2 and CH4 flux relative to corn but were not significantly
different from each other (Tables 2.2 and 2.3). The effects of corn and soybean crops on
N20 flux were not significantly different fi'om each other, but wheat increased N20 flux
in comparison to corn (Table 2.4). The coefficients for the effect of crops reveal a
consistent order of effect such that each crop increases gas flux in the order wheat >
soybean > corn. Timing of fertilization did not have a significant effect on any of the
gases modeled.
Historical Management Factors

Of the five historical management factors tested, no single factor in the CO2 and
CH4 models had a significant effect in all four soil treatments. However, C/N ratio had a
significant effect in all but the historical agriculture treatment in the CO2 models and the
same was true of percent soil moisture in the CH4 models. Percent soil moisture was also
the factor with a significant effect in the most treatments for the N20 models, with
significant effects in all four treatments. The N20 and CH4 models were similar in that
the deciduous forest had the greatest number of significant historical management factors
of all treatments, while the never tilled treatment had the most significant factors in the

CO2 model.

28

.83 m e 285%.. $505.53 .2. so; aEeEooo 85 as? as g 2F _.
duh—8a £5 .80 8:83 05 an vow: 83 no.5 E8 25. 43:? v5 £56.38 .53 3952. .6382 .353 5 co eon—no.0 & 2.0 o
.85 w a a 8:85.53 .:.... 2a .85 w a 3 8:85am: .:.... .2... w a a 8:85.? £82 .:..
< .8892. ac 83% a a? 33:8 203 "5 ea Hz .5 .88»: co $938 2 .23 33225 .53 as B 2e 6.. 226. sewage»; m

 

 

 

 

.Bmoficea E :03
0.8 953508 .5.“ tote 23.—Sm 38¢ a E omega E5 25 :08 com A7.» N.:. ~00 mac mace—«>350 ~00 E omega 05 2338 £56530 .
mum Pm Sn «9. 88¢ng .«o h3:52
86 86 86 86 c A m no.5
and and and and £3 noun—agon— 20 £80
.:. 5.8 42 : ad «3. . one 5.... 8.: 82 .3235 .25
:3 A89 N26 .3 28.8 ”:6 .2... 23.8 2 _.¢ 8: A89 :86 BERKEob a: 83532
18.8086 2:. 23.8 mafia 3.. $5.8 v26 .:.... 28.8 head 0.5%th a? 85:83.2
£98 :56. .:.... G98 2 _.c 2. 28.8 23.: :98 «56 $23 C 83362.—
.. 20.8 Sod _. G98 89? 398 Bed 298 Sod. 6.23 8 now—3360.5 aaogohém
as. 3.0.8 :20. a: An .60 036.. 32.8 086 a 22.8 98.? 20
A: .8 ecmd. 23.8 and .:.... $5.8 ewN Ante 3 «.6 89.30 c\.
28.8 «mod 9... 3°60 mid. 9:60 mid. 23.8 cad 5388880 83:95.2
$5.8 936. :5 $08 m 36 o 32.8 mend. 28.8 «56. 8285:0230 32:2 8080352
.:.... 28.8 :30 a 23.8 286 96.8 Sod $6.8 v36 8:632 now ..\o 33.8%:
- - - 898 :56. cog—Eek
. u .. a .99 Adev mn.— no.8 30:3 acoaow
- - - u :3 22.8 She no.8 Sow 80> .550
38o..— 33280 3:2. .962 2.53% .353: o Each—Mat 139.930 88$ bouofio 88a..—
3. 4.86880 BEE."—

 

ssa 60 £2.33 e38". .6 an»; Sansone 032:2 “Na 2...;

29

.8. .c N n. .gt... 2.585%... .9. 89.» 8.86508 no.6 82.3 ...... 58.38 on... ..

.m.mb§m.5.o.o£_omzofim¢oom:§no.o§oofi u§3v§d§§.§§3§_§§§8vonnobflh0u

...... w .. ... 853...? .:.... ...... .8... w .. ... 8.553.. ...... .2... .v. .. ... 880......» 8.8%... ....

< .5802. .... 82w»... a ....a 83......» 203 .o ...... ...z ..E €on.. ... 89...... ~. 5.3 33.2.... 22., 5.... ...o 2.. 3. 32.... 8:85.35.»

68055.3 ... =o>

2.. 8.8.0503 ..o.. 95.5. gnaw 8.8.. a a. 098.... ...... 25 no.8 8.. A...» ~.E ~00 w. ace—«>33 ~00 ... vac—Eu o... 8:82.. 35.0580 .

 

 

 

 

m- E .2 ...... aouagno... 2.5.2
...... 8... 8... 8... .. A .. .2.
a... ...... 8... ...... m... 8.3.58.8 .o ....80
... .3. «...... ...: ...... 3w ... .3. 2... a... a... .823... .28
... .... .... Sn... .3... “8.... .. 8...... v: .... 3...... «.8... gas... .... 35......
.... .... m~. ... .-.... ..-.... .8... n8... 3...... 3...... 25836»... ..< 58.3.
.2. .8... 3.. ... 3...... 3.... : .8... 8.... . .8... .8... .2... ... 8.3.3.2..
... .3... ma... 3...... a3... .3... 3...... .8... ...... .2... n. 8.9%.... 3858......
.. 8n... 5.... .3... ... ~... .3... an... 8.... ~23. 2.0
a... 2...... 39...... ... 8:8... ... ...... 2.... 8.3.x.
.. an... 2..- .3... .:.... .8... ....~... .3... 2.... ......EcooSo $206.5.
.. .3... g... 3...... .18....56. Goes... ......S..88o 3.3.2 308035:
... 6.... SH... ... ...... ... n... 3...... ...... . 3...... ~z... 23...... =8 ... .85....
- - - 2...... ...... 83.20..
.. .. u .. :3 3...: find no.6 .8; 3080»: =. _
- - - u 2.. :68 mm. no.8 Sow 80> 32.50
.83... 32....8o 8.... .962 2330.9. .882... 0 25.3.9. 18:8..80 .88.. beams“. .88..
.... 38.0530 8.25....

 

.3... £0 Essa... 838.. ... are... 8.853. 0......3... .3 0......

30

.3... w .. .88.... 8.82.28... 22: 98.0.28. 88 .8.>9 .8 82.8 2F _.

8.3.3:. 25...... 22.8.... ofiaggnEoEooofi .2823 .28 823923 885033.. 5.8.2 .222... .... Egafihoo

...... w .. ... 8.828.. .:.... ...... .8... w .. ... 8.828.. ...... ...... w .. a 8.828.... 8.82.... ......

< .888... .... 80.32. a 5.3 3.8:—«>0 203 ..E .2... ....Z ...: 8.2.8.... .... 8950.. N. 5.3 3.3.2,... 203 8.... .5 2.. ...... 20>... 8:855.m.m

 

 

 

 

80855.2... ... 20>
98 3.5.0503 ...... 82.... 98.28% .38.. a ... «wad—.0 ...... 2... :80 ...... A...» NE 50 3 8:23.30 .00 ... 09.2.0 2.. 838:. 35.0530 .

8. n8 .8 ...... 88:88... .382
...... ...... ...... ...... .. A .. 8...
...... 8... ...... 8.... .N2. 88558.8... 280

... .2. En- .... at- .... .. 3...... ...... n..- .822... 8.5
.8... 82.. . .8... E... ... .8... 8.... . .... 8... 2388...... ..< 88......
.8... 8.... .8... 8.... .2 .... 8. ... ... .8... $8 28888.... ..< 8.8.32
.2 .... 3...... .:. .8... N8... .8... 2...... . .8... 8.... .....8 h. 8.88.8...

: .2... .8... : .8... 8...... .2... 3...... : .3... 8.. .88 .... 8.8.8.8... 88.5.2.8...
.. .8... 8.. .3... an... .3... .8... ...... .:.... 2...
.3. 8.... ....m. 8..- ...m. ...... .8. N8- 8.80...
...... N...” .8... 8.... .8... 2 .. .8... .8... 8.8.888 8.85:2

.. .8... t. .5. :8 .8... 2...... . .8... 8.. 8.88880 38.2 .8888...

... .8... 8.. : .8... 8.... ...: .2 .... m3... ... .8... ...... 28...: =8 ... .88....
- - - .2 .... 2.... 8.8.2....

.. - - _. .. $3 me no.6 .82? 2.080352

- - - u And. a...— aob gem 28> 2.2.50

.88.. 8828.. 8.... .952 8.8.8.. .88.»... o 238...? 28.880 8.8... Eu»... 8.8...

.. 8.0.0280 828......

 

...... 0.2 8882.... 8.8.. ... 828... 8.328.. 0.832 ....N 2...:

31

Environmental Factors

There were four environmental factors tested and in the C02 model, maximum
and minimum temperatures were significant in all soil treatments. C02 flux increased as
either air temperature measurement increased. Each treatment had a minimum of two
factors with a significant effect on C02, and all four factors were significant in the never
tilled treatment. No single environmental factor was significant in the N20 or CH4
models for all soil treatments. However, in the N20 model, cumulative precipitation after
three days and minimum temperature were significant in three out of four treatments
making them the factors with the most consistent impact. In the CH4 model, this was true
of cumulative precipitation after seven days. The conventional agriculture site had the
most significant factors of the N20 models, while the deciduous forest had the most in the
CH4 models.
Recovery from Agriculture

The historical agriculture treatment was compared to the other successional
treatments (NT and DF) and the conventional agriculture treatment to determine if factors
affecting gas flux from plots abandoned fi'om agriculture were more similar to those
afl‘ecting gas fluxes in mid- to late-successional or conventional agricultural treatments.
In the case of CH4 and C02 flux, the historical agriculture site had more variables in
common with the successional treatments than with the conventional agricultural
treatment. This may indicate that the microbial communities in the sites abandoned fiom
agriculture are more similar to those in the successional treatments than those in sites

currently used for agriculture. In the N20 models, the historical agriculture treatment had

32

only two variables that were significant, and these were also significant in the
successional and agricultural treatments.
Overall Modeling Results

Of the three gas fluxes, C02 consistently had the highest coefficient of
determination for each of the four treatments (R2 = 0.29 to 0.50). In addition, when
looking at all factors, all except fertilization had a significant effect in at least one soil
treatment, resulting in C02 being the gas flux with the largest number of significant
variables. The N20 and CH4 models each could only explain a small proportion of the
variation in the data (R2 = 0.08 to 0.19 and R2 = 0.09 to 0.33, respectively). The
deciduous forest had the highest R2 value for each of the three gases when compared to
the other treatments (R2 = 0.19 to 0.50). The intercept was commonly a significant
variable in the historical agriculture, never tilled and deciduous forest treatments. This
indicates that there are unmeasured parameters in these treatments that are responsible for
explaining some of the gas flux from these sites.

The impact of measuring moisture by percent soil moisture or WFPS was also
investigated (data not shown). In most cases, percent soil moisture was significantly
different from zero, whereas WFPS was not, indicating percent soil moisture is a better
variable to depict soil moisture in our models. Use of percent soil moisture instead of
WFPS was found to sharpen and clarify the relationship found between soil moisture and

gas fluxes, therefore all subsequent analyses included only percent soil moisture.

33

DISCUSSION:

The extensive worldwide acreage converted to agricultural uses has contributed to
an increased concentration of greenhouse gases in Earth’s atmosphere. The annual net
release of C02 fiom agriculture is estimated to contribute ca. 14% of current fossil firel
emissions [1]. Agricultural practices also increase the flux of N20 relative to non-
agricultural soils [4, 5], and decrease the capacity of soil to serve as a methane sink.
Conversion of tropical, subtropical and temperate soils to agricultural use is estimated to
have decreased the methane sink by 3 to 9% worldwide [14, 15]. To identify trends in
the variability of greenhouse gas fluxes from soils in agroecosystems associated with
different crops and under different management regimes at the KBS LTER we used a
linear production function approach.

The coefficients of determination for these analyses were generally lower than
those obtained in previously published studies (Table 2.5), which is due to experimental
design and length of the studies. Many of the higher coefficients of determination (R2)
were obtained from experiments in which a single independent variable was altered, and
so the likelihood of explaining that data is greater. Also contributing to their increased
predictive power is the fact that the majority of the studies took place over a shorter time
period, eliminating annual variations from the model. In a meta-analysis of studies on
N20 flux published between 1980 and 1997, Sozanska and colleagues [16] found that
only 3% of all studies were longer than one year and had measurements taken on at least
a weekly basis. While process modeling is the usual approach to predict gas fluxes,
empirical models such as ours are useful for identifying variables important in describing

flux and calculating how well those variables account for the flux observed.

34

.bco 8:. 35:8 .5.» a

 

 

ban—am
comp—z: v5 ._o>o_ 533593 €2.33 83 538%? <mD
as .60 .835 ..anza .5: sea as 8: .8» _ a a§§§c§§a 35m
$39» m ..o>o :83
:3 eaeaefi 35595305 «5.55 a 3.: <3 .32.: $.8an
<mD
3 : 2382.th 2:58 e we .52 E8 828 .2304 NOD
Amado» m .63 :83
83 282 ficoaflsmwoav 8.003 R a 2. meagowbmfleu 33am
vac—wcgmea
a: 282 2:58 2 a an 30:202. 888th
<mD
a : eaeaeoe mass m. s .52 as e28 233 £0
mega .Esése 82 . 82 Eat 85:95 335%.; .8339
G: 533% 2:55 ...oeflzeon 38% masses: 3. -8: a. .ea_3_e<
v§_=_m\m=8
r : 252 e8» m 8 2.8 58. Ea as 328 oaz
00:20qu Bea—=35: beam we 538.5 4.x.v noggin—089 new; :8 “6.5302
baccaraxm use?» .8 256580 80

 

moan—E 30 0305680 maven: Eowoam 058.630 3 £352 command 032:2 wfimD 838m "Wm ...—cap.

35

In our analysis, there was consistently a greater ability to explain the variability in
C02 flux than that of CH. or N20, and the same factors that explained variability in CO;
flux were consistently less effective in accounting for variability in N20 or CH4 flux.
Carbon dioxide flux had a greater proportion of environmental factors that were
significant than the N20 or CH4 models. The differences in R2 values are interpreted to
be due to differences in the microbial metabolisms involved in the production of each
gas. Carbon dioxide is produced by most microbes in soil: it is a product of fermentation
as well as both aerobic and anaerobic respiration. However, CH4 consumption and N20
production are the result of specialized forms of microbial metabolism that are favored
under specific conditions, and the microbes that carry out these reactions are less
abundant in soil. Methane consumption is carried out under aerobic conditions by
methanotrophic bacteria, while N20 production is favored under anaerobic conditions
and is likely produced primarily by denitrifying bacteria. The organisms carrying out
these specialized metabolisms make up a very small proportion of the bacterial soil
population. Estimates of nitrifier abundance place them at a maximum of 0.01% of the
total bacterial population in soils from the KBS LTER [23]. Soil denitrifier and
methanotroph abundance has been shown to be approximately 0.01% [24, 25] and 0.3%
[26] respectively, of the total bacterial population. Therefore, changes in variables that
have a general effect on cellular metabolism would be expected to have a greater effect
on rates of C02 production.

In the agricultural treatment, there were consistent differences in the effect of
each crop on gas flux. Wheat was associated with the highest average flux of N20 and

CO; and the lowest ability to oxidize CH4. The three crops grown at KBS vary in their

36

amount and type of root exudate [27], plant carbon content and timing of carbon
availability [28], as well as the timing of fertilization. These differences affect the
quantity and quality of nutrients microbes require; therefore it is possible that the annual
rotation of crops at KBS results in annual changes in microbial communities that vary in
their metabolic capabilities, resulting in differences in gas fluxes. Recent studies have
shown that different species of plants will influence the composition of microbial
communities, thereby changing the metabolic capabilities of the community [29-33].
This result suggests that a possible mitigation option for greenhouse gases would be
through the growth of only certain crops.

Another notable observation in the agricultural treatment was the lack of a
significant effect of the timing of fertilization on fluxes of N20 and CH4. Nitrogenous
fertilizer would be expected to cause an increase in N20 flux due to an increase in
nitrogen available to those organisms producing it, and a decrease in CH4 consumption
due to the inhibitory effects of ammonium on CH4 oxidation. Despite the lack of a
significant effect from the timing of fertilization, nitrate concentration was found to have
an effect on N20 flux, which has been found in other studies. Syvalso and colleagues
[17] found no correlation between N20 flux and the amount of fertilizer applied, but did
identify a correlation with the total amount of nitrogen in soil. In addition, a previous
study at the KBS LTER concluded that it is nitrogen availability and not fertilizer
specifically that leads to increased N20 flux [5]. Nitrogen from fertilizer will not become
available to N20-producing organisms until the proper conditions develop. This results
in a disconnect between the timing of N20 production and fertilizer application. Future

modeling efforts, especially for N20, may be better served by including measures on

37

nitrogen availability at the time gas measurements are taken in addition to amounts of
fertilizer applied.

The insignificant effect of fertilization on CH4 flux is likely indicative of a long-
term response of the microbial community to fertilization. After years of fertilization,
CH4 flux does not respond to subsequent N input, and the lack of response is likely due to
the selection of a methanotroph community that is unaffected by fertilization and other
agricultural perturbations [34]. The possibility of such a community shift has rarely been
addressed, but a study characterizing the methanotroph communities in paired
agricultural, successional and forest soils found that agricultural soil only had a subset of
the methanotroph groups in the forest soil [35]. Other characterizations of agricultural
soil without paired sites have also found that they lack sometime dominant forest
methanotrophs [3 6]. Therefore, a methanotroph community shift driven by long-term
agricultural perturbations is likely to have occurred at KBS.

Our data suggests that this methanotroph community that is unaffected by
agricultural perturbations is also largely unresponsive to changes in the environmental
variables we have modeled, and can be thought of as being perturbation resistant. The
lack of significant variables and low CH4 R2 values observed in the conventional
agriculture, historical agriculture and never tilled sites imply those treatments are likely
dominated by these seemingly perturbation resistant methanotrophs. The comparatively
high R2 value and multiple significant variables found for the deciduous forest indicates
that perturbation sensitive methanotrophs are likely present at significant levels. We
hypothesize that it is these perturbation sensitive methanotrophs that lead to the model’s

greater explanatory power in the deciduous forest.

38

It is not obvious if similar arguments apply to the flux of N20 and C02, but
studies suggest that the transition from native lands to agriculture causes a shift in the
C02 producing microbial community [3 7, 38]. The authors of these studies hypothesize
that changes in labile organic matter that accompany the land-use transition is the driving
force behind the differences in the C02 producing microbial community. In addition,
Stevenson and colleagues [39] found distinct functional patterns in the C02 producing
microbial community on a landscape scale between pasture and forest soil. Presumably
the differences were caused by variation in litter input quality and management practices.
Such explanations would be supported by our data since a C02 producing microbial
community with a comparatively high R2 value is found in the forest, perhaps due to the
presence of more labile organic matter, land management practices, and different litter
input.

While the above explanations of variability in gas flux are consistent with
expectations based on microbial physiology and ecology, it is of note that there are other
differences between the treatments that may also influence the flux of greenhouse gases.
As discussed previously, the above-ground plant community differs among treatments
and is likely to impact the microbial community. In addition, primary productivity and
pH also change between treatments. None of these factors were accounted for in our
model, but like labile organic matter and the C02 producing microbial community, may
impact the composition and activity of the microbial community and the associated gas
fluxes.

The linear production approach implemented in this study provided the important

insight that, especially for N20 and CH4, the microbial community needs to be assessed

39

in order to build an accurate process model. Process models typically assume that
environmental and land management factors are the primary determinants of soil gas
flux. Process models tend to treat microbial communities in soil as a “black box,” with
associated inputs and outputs that are influenced by a collection of environmental
parameters. Measures of the composition or diversity of microbial communities are
generally not included in models, both because of the difficulty in collecting this data and
the general expectation that measurements of the precise composition of the community
are not necessary due to metabolic redundancy. Recent studies however, suggest that the
diversity and composition of microbial communities may help to explain differences in
ecosystem functions [40—44].

The findings presented in this study are relevant to both current year and
historical management plans given the economic costs associated with loss of fertilizer
and climate change brought about by increased concentrations of greenhouse gases.
Climate change has been implicated in contributing to increased temperatures, altered
precipitation patterns and the increased frequency of extreme weather events. These
factors all serve to depress crop yields and increase production risks so that the gap
between food production and consumption between developed and developing nations is
predicted to increase [45]. The recent increase in weather-related catastrophes has also
had effects on regional and national economies, with windstorrns from 1983 to 1992
alone accounting for $88.1 billion in economic losses [46]. Given the impact of
greenhouse gas emissions, it is important to begin to treat microbial communities,
especially methanotrophs and denitrifiers, as important factors in modeling variations in

gas flux.

40

CONCLUSION:

Land management practices and environmental factors influence the exchange of
greenhouse gases between soils and the atmosphere by altering the metabolism of
microbial communities in soil. It is concluded that since C02 is a byproduct of all types
of heterotrophic respiration in soil, its flux will be responsive to factors that enhance
microbial metabolism in general, and that this phenomenon explains the high coefficients
of detemrination for explaining variation in CO; flux at the KBS LTER. Conversely,
since the consumption of CH4 and production of N20 require the activity of microbes
with specialized metabolic pathways, environmental parameters alone explained a
smaller percentage of the flux of these gases. Advances in the modeling of CH4
consumption and N20 production will likely require estimates of the diversity and

activity of the microbial populations that oxidize CH4 and emit N20.

41

REFERENCES:

l.

10.

11.

Schlesinger, W.H., An overview of the carbon cycle, in Soils and Global Change,
B.A. Stewart, Editor. 1995, CRC Press, Inc.: Boca Raton. p. 9-25.

Murty, D., M.U.F. Kirschbaum, R.E. McMurtrie, and A. McGilvray, Does
conversion of forest to agricultural land change soil carbon and nitrogen? a
review of the literature. Global Change Biology, 2002. 8(2): p. 105-123.

Smith, K.A., K.E. Dobbie, B.C. Ball, L.R. Bakken, B.K. Sitaula, S. Hansen, R.
Brurnme, W. Borken, S. Christensen, A. Priemc, D. Fowler, J.A. Macdonald, U.
Skiba, L. Klemedtsson, A. Kasimir-Klemedtsson, A. Degorska, and P. Orlanski,
Oxidation of atmospheric methane in Northern European soils, comparison with

other ecosystems, and uncertainties in the global terrestrial sink. Global Change
Biology, 2000. 6(7): p. 791-803.

Granli, T. and QC. Bockman, Nitrous oxide from agriculture. Norwegian Journal
of Agricultural Sciences Supplement, 1994. 12: p. 7-128.

Robertson, G.P., B.A. Paul, and RR. Harwood, Greenhouse gases in intensive
agriculture: contributions of individual gases to the radiative forcing of the
atmosphere. Science, 2000. 289: p. 1922-1925.

Parton, W.J., A.R. Mosier, D.S. Ojima, D.W. Valentine, D.S. Schimel, K. Weier,
and A.E. Kulrnala, Generalized model for N 2 and N20 production fiom
nitrification and denitnfication. Global Biogeochemical Cycles, 1996. 10(3): p.
401-412.

Parton, W.J., M. Hartman, D. Ojima, and D. Schimel, DA YCENT and its land
surface submodel: description and testing. Global and Planetary Change, 1998.
1904): p. 3548.

Li, C.S., S. Frolking, and TA. Frolking, A Model of Nitrous-Oxide Evolution
fi'om Soil Driven by Rainfall Events: 1. Model Structure and Sensitivity. Journal
of Geophysical Research-Atmospheres, 1992. 97(D9): p. 9759-9776.

IPCC, Revised 1996 Guidelines for National Greenhouse Gas Inventories. 1997,
Intergovernmental Panel on Climate Change.

Elliott, E.T., J.W. Heil, E.F. Kelly, and H.C. Monger, Soil structural and other
physical properties, in Standard soil methods for long-term ecological research,
P. Sollins, Editor. 1999, Oxford University Press: New York. p. 74-85.

Jarrell, W.M., D.E. Armstrong, D.F. Grigal, E.F. Kelly, H.C. Monger, and D.A.
Wedin, Soil water and temperature status, in Standard Soil Methods for Long-

42

12.

l3.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Term Ecological Research, P. Sollins, Editor. 1999, Oxford University Press:
New York. p. 55-73.

Robertson, G.P., P. Sollins, B.G. Ellis, and K. Lajtha, Exchangeable ions, pH, and
cation exchange capacity, in Standard Soil Methods for Long- Term Ecological
Research, P. Sollins, Editor. 1999, Oxford University Press: New York. p. 106-
l 14.

Sollins, P., C. Glassman, B.A. Paul, C. Swanston, K. Lajtha, J.W. Heil, and ET.
Elliott, Soil carbon and nitrogen: Pools and fractions, in Standard Soil Methods
for Long- Term Ecological Research, G.P. Robertson, Editor. 1999, Oxford
University Press: New York. p. 89-105.

Breas, 0., C. Guillou, F. Reniero, and E. Wada, The global methane cycle:
Isotopes and mixing ratios, sources and sinks. Isotopes in Environmental and
Health Studies, 2002. 37(4): p. 257-379.

Wuebbles, DJ. and K. Hayhoe, Atmospheric methane and global change. Earth-
Science Reviews, 2002. 57(3-4): p. 177-210.

Sozanska, M., U. Skiba, and S. Metcalfe, Developing an inventory of N20
emissions from British soils. Atmospheric Environment, 2002. 36(6): p. 987-998.

Syvasalo, E., K. Regina, M. Pihlatie, and M. Esala, Emissions of nitrous oxide
from boreal agricultural clay and loamy sand soils. Nutrient Cycling in
Agroecosystems, 2004. 69(2): p. 155-165.

Peterjohn, W.T., J.M. Melillo, P.A. Steudler, K.M. Newkirk, RP. Bowles, and
JD. Aber, Responses of trace gas fluxes and N availability to experimentally
elevated soil temperatures. Ecological Applications, 1994. 4(3): p. 617-625.

Bradford, M.A., P.A. Wookey, P. Ineson, and H.M. Lappin-Scott, Controlling
factors and eflects of chronic nitrogen and sulphur deposition on methane
oxidation in a temperate forest soil. Soil Biology & Biochemistry, 2001. 33(1): p.
93-102.

Steinkamp, R., K. Butterbach-Bahl, and H. Papen, Methane oxidation by soils of
an N limited and N fertilized spruce forest in the Black Forest, Germany. Soil
Biology & Biochemistry, 2001. 33(2): p. 145-153.

Rustad, LE. and LI. Fernandez, Experimental soil warming efl'ects on C02 and
CH4 flux from a low elevation spruce-fir forest soil in Maine, USA. Global
Change Biology, 1998. 4(6): p. 597-605.

Craine, J.M. and D.A. Wedin, Determinants of growing season soil C02/lax in a
Minnesota grassland. Biogeochemistry, 2002. 59(3): p. 303-313.

43

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Phillips, C.J., D. Harris, S.L. Dollhopf, K.L. Gross, J.I. Prosser, and B.A. Paul,
Efi’ects of agronomic treatments on structure and junction of ammonia-oxidizing
communities. Applied and Environmental Microbiology, 2000. 66(12): p. 5410-
5418.

Martin, K., L.L. Parsons, R.E. Murray, and MS. Smith, Dynamics of Soil
Denitrrfier Populations - Relationships between Enzyme-Activity, Most-Probable-
Number Counts, and Actual N-Gas Loss. Applied and Environmental
Microbiology, 1988. 54(11): p. 2711-2716.

Henry, S., E. Baudoin, J .C. Lopez-Gutierrez, F. Martin-Laurent, A. Baumann, and
L. Philippot, Quantification of denitrifying bacteria in soils by nirK gene targeted
real-time PCR. Journal of Microbiological Methods, 2004. 59(3): p. 327-335.

Kolb, S., A. Carbrera, C. Karnmann, P. Kampfer, R. Conrad, and U. Jackel,
Quantitative impact of C02 enriched atmosphere on abundances of

methanotrophic bacteria in a meadow soil. Biology and Fertility of Soils, 2005.
41(5): p. 337-342.

Jones, D.L., Organic acids in the rhizosphere - a critical review. Plant and Soil,
1998. 205(1): p. 25-44.

Buyanovsky, GA. and G.H. Wagner, Soil respiration and carbon dynamics in
parallel native and cultivated ecosystems, in Advances in Soil Science: Soils and
Global Change, B.A. Stewart, Editor. 1995, CRC Press, Inc.: Boca Raton. p. 209-
217.

Nehl, D.B., S.J. Allen, and J .F . Brown, Deleterious rhizosphere bacteria: An
integrating perspective. Applied Soil Ecology, 1997. 5(1): p. 1-20.

Grayston, S.J., S.Q. Wang, C.D. Campbell, and A.C. Edwards, Selective influence
of plant species on microbial diversity in the rhizosphere. Soil Biology &
Biochemistry, 1998. 30(3): p. 369-378.

Siciliano, S.D., C.M. Theoret, J.R. de Freitas, P.J. Hucl, and 1.1. Germida,
Dtfl'erences in the microbial communities associated with the roots of drflerent
cultivars of canola and wheat. Canadian Journal of Microbiology, 1998. 44(9): p.
844-851.

Wieland, G., R. Neumann, and H. Backhaus, Variation of microbial communities
in soil, rhizosphere, and rhizoplane in response to crop species, soil type, and
crop development. Applied and Environmental Microbiology, 2001. 67(12): p.
5849-5 854.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

Kent, AD. and E.W. Triplett, Microbial communities and their interactions in
soil and rhizosphere ecosystems. Annual Review of Microbiology, 2002. 56: p.
21 1-236.

Suwanwaree, P. and G.P. Robertson, Methane oxidation in forest, successional,
and no—till agricultural ecosystems: Eflects of nitrogen and soil disturbance. Soil
Science Society of America Journal, 2005. 69(6): p. 1722-1729.

Knief, C., S. Vanitchung, N.W. Harvey, R. Conrad, P.F. Dunfield, and A.
Chidthaisong, Diversity of methanotrophic bacteria in tropical upland soils under
drflerent land uses. Applied and Environmental Microbiology, 2005. 71(7): p.
3826-3831.

Knief, C., A. Lipski, and RF. Dunfield, Diversity and Activity of Methanotrophic
Bacteria in Diflerent Upland Soils. Appl. Environ. Microbiol., 2003. 69(11): p.
6703-6714.

Degens, B.P., L.A. Schipper, G.P. Sparling, and M. Vojvodic-Vukovic,
Decreases in organic C reserves in soils can reduce the catabolic diversity of soil
microbial communities. Soil Biology & Biochemistry, 2000. 32(2): p. 189-196.

Schipper, L.A., B.P. Degens, G.P. Sparling, and LC. Duncan, Changes in
microbial heterotrophic diversity along five plant successional sequences. Soil
Biology & Biochemistry, 2001. 33(15): p. 2093-2103.

Stevenson, B.A., G.P. Sparling, L.A. Schipper, B.P. Degens, and LC. Duncan,
Pasture and forest soil microbial communities show distinct patterns in their
catabolic respiration responses at a landscape scale. Soil Biology &

Biochemistry, 2004. 36(1): p. 49-55.

Cavigelli, M.A. and G.P. Robertson, The functional significance of denitrrfier
community composition in a terrestrial ecosystem. Ecology, 2000. 81(5): p. 1402-
1414.

Holtan-Hartwig, L., P. Dorsch, and LR. Bakken, Comparison of denitrifying
communities in organic soils: kinetics of N0; and N20 reduction. Soil Biology &
Biochemistry, 2000. 32(6): p. 833-843.

Cavigelli, M.A. and G.P. Robertson, Role of denitrrfier diversity in rates of
nitrous oxide consumption in a terrestrial ecosystem. Soil Biology and
Biochemistry, 2001. 33: p. 297-310.

Rich, J.J., R.S. Heichen, P.J. Bottomley, K. Cromack, Jr., and DD. Myrold,
Community Composition and Functioning of Denitrijfying Bacteria fiom Adjacent
Meadow and Forest Soils. Applied and Environmental Microbiology, 2003.
69(10): p. 5974-5982.

45

45.

46.

Webster, G., T.M. Embley, T.E. Freitag, Z. Smith, and J .I. Prosser, Links between
ammonia oxidizer species composition, fimctional diversity and nitrification
kinetics in grassland soils. Environmental Microbiology, 2005. 7(5): p. 676-684.

Fischer, G., M. Shah, F .N. Tubiello, and H. van Velhuizen, Socio-economic and
climate change impacts on agriculture: an integrated assessment, 1990-2080.
Philosophical Transactions of the Royal Society B-Biological Sciences, 2005.
360(1463): p. 2067-2083.

Tucker, M., Climate change and the insurance industry: the cost of increased risk
and the impetus for action. Ecological Economics, 1997. 22(2): p. 85-96.

CHAPTER 3

COMPOSITION, DIVERSITY, AND N20 FLUX OF DENITRIFIERS IN AN
AGROECOSYSTEM

ABSTRACT:

Denitrifying rrricroorganisms in soil act as a source and sink of nitrous oxide and
contribute to losses of nitrogen from agricultural fields. The denitrifier community
composition and diversity of three soil treatments varying in level of agricultural
management and N20 flux were compared. Comparisons of proteobacterial nirK
sequence libraries indicated that community composition of successional sites abandoned
fi'om agriculture 15 years ago still showed impacts from agriculture, and that the
community composition of the 0-7 cm portion of this site had diverged from that found at
13-20 cm. Productivity (%C) did not have a significant relationship with measures of
denitrifier diversity, but an analysis using Spearrnan’s Correlation Coefficient indicated a
weak, positive relationship between relative disturbance and diversity (P = 0.1 to 0.2). In
addition, a significant, positive relationship between ribosomal RNA (rrn) operon copy
number and growth rate of dissimilatory nitrate reducers was found, indicating that
denitrifiers with low and high rm operon copy numbers differ in their ecological
strategies. There was a significant effect of agricultural treatment on evenness (P =
0.0476), with the conventional agriculture treatment at 0-7 cm having the highest
evenness. This, along with rank/abundance curves, indicates that there are nirK OTU’s
in the never tilled and historically tilled treatments that may play an important role in

minimizing the amount of N20 emitted from sites currently not used for agriculture.

47

INTRODUCTION:

Denitrification is a dissimilatory process carried out by a phylogenetically diverse
group of bacteria and fungi that are predominantly heterotrophic and facultatively
anaerobic [1]. During denitrification, nitrate is reduced in a step-wise manner to nitrite,
nitric oxide, nitrous oxide, and finally dinitrogen gas [1, 2]. Typically, denitrification
results in the release of inorganic nitrogen as either N20 or N2, and the N20 mole ratio
and N20 flux produced by soil denitrifiers varies based on physical properties such as soil
pH, oxygen status, available carbon, or soil moisture history [3-5]. Knowledge of the
factors influencing N20 mole ratio and flux is important, as N20 is a greenhouse gas that
contributes to global warming and destroys ozone. Natural sources of N20 contribute
approximately 10 Tg N - yr’l to the global budget, and of this, soil contributes 65% [6].
In addition, in agroecosystems loss of nitrogenous fertilizer as gaseous products decrease
farm profit margins as greater amounts of fertilizer are used to balance the loss of
inorganic nitrogen fi‘om the field [7].

While the physical factors affecting N20 flux have received much attention, there
is also evidence that differences in denitrifier communities can influence N20 flux [8-12].
Traditionally, microbial diversity especially within groups carrying out a specific
biogeochemical process, was assumed to be functionally redundant. This has led to the
creation of many process models that predict N20 flux, but do not account for functional
differences in community composition. For this reason, many models estimating N20
flux from agricultural soils remain limited [13]. For example, IPCC guidelines [14]

consider all agricultural systems to be equivalent, which does not take into account

48

environmental factors affecting the microbial communities that produce and remove N20
from the soil environment.

Determining factors that influence or maintain diversity have been long-time
goals of ecological studies. Two factors found to be important in both macro- and
microbiological communities are disturbance and productivity [15-18]. Most studies on
disturbance find that communities follow the predictions of the Intermediate Disturbance
Hypothesis (IDH), which states that species diversity will be highest in environments
having an intermediate level of disturbance [19]. The proposed differences in diversity
are due to a balance between reduced richness fi'om competitive exclusion at low
disturbance levels and from increased mortality at high disturbance levels [19, 20].
Productivity of an environment also has effects on diversity; however this relationship is
less predictable than that found with disturbance. It can vary with the directness of the
productivity measure, trophic level of the organism(s) being studied, or the scale of the
experiment [21, 22].

In addition to understanding factors that influence community composition and
diversity, it is important to understand the underlying ecological strategies employed by
organisms. It has been demonstrated that the ribosomal RNA (rrn) operon copy number
of aerobic, heterotrophic bacteria reflects their ecological strategy [23]. In experiments,
high copy number organisms were able to respond quickly to new resources while low
copy number organisms grew more slowly. This is thought to be the result of a trade-off
occurring between benefits incurred fi'om being able to produce ribosomes quickly when
resources are available versus the high energetic cost of maintaining many ribosomes in a

low nutrient environment. Agricultural treatments at the KBS LTER vary in the

49

availability of nutrients and frequency of nutrient inputs, therefore different ecological
strategies may be favored in different treatments. To determine specifically whether the
rrn operon copy number of organisms might reflect an ecological strategy, the
relationship between growth rate, phylogeny and rm operon copy number was explored.
Dissirrrilatory nitrate reducers capable of either denitrification or reduction of nitrate to
ammonia were included in this portion of the study in order to determine how widespread
the potential trade-off was amongst organisms that carrying out this type of metabolism.

Many studies assessing microbial diversity or community composition rely on the
analysis of 168 rDNA genes. Denitrifying microorganisms belong to a phylogenetically
diverse group; therefore it is not possible to design 16S rDNA primers that would capture
sequences only from organisms with denitrifying capability. Instead, a functional gene in
the denitrification pathway was used to assess the denitrifier community. Nitrite
reductase (N ir) catalyzes the key step in denitrification since it produces the first gaseous
product in the pathway and is responsible for the first step in the pathway that
distinguishes between nitrate respirers and denitrifiers [1]. There are two types of nitrite
reductase genes that encode for enzymes that are structurally different, but functionally
the same. NirK encodes for a copper containing nitrite reductase and nirS encodes for a
cytochrome cdl containing enzyme. Known denitrifiers possess either one of these
genes, but not both [24, 25]. The work presented here indicated that nirS was either not
present or in low abundance in the soil treatments studied, therefore diversity was
assessed using nirK nucleotide sequences.

One of the primary goals of this study was to determine whether proteobacterial

denitrifier community composition differed between sites that differ in annual average

50

N20 flux and history of agriculture. Sampling was performed at the Kellogg Biological
Station Long-Term Ecological Research (KBS LTER) site where average N20 flux from
a conventional agricultural treatment (3.22 :l: 0.45 g N20-N-ha’l°d") is approximately
three times higher than fiom either a historically tilled treatment (0.92 :1: 0.08 g N20-
N'ha"-d'l) or a mid-successional, never-tilled treatment (1.13 :1: 0.11 g N20-N-ha‘l'd")
[26]. Measurements of greenhouse gases taken at the KBS LTER indicate that N20
production from the annually cropped treatments is the major contributor to global
warming potential (GWP) [27]. Other primary goals were to determine whether there
were differences in the diversity of denitrifier communities, to identify factors
contributing to differences in diversity and composition, and finally, to determine
whether differences were associated with changes in N20 flux. Secondary goals included
an exploration of dissimilatory nitrate reducer ecological strategy, comparison of
replicate clone libraries, and an assessment of the impact of agriculture on denitrifier

communities.

MATERIALS & METHODS:
Study Site 8: Soil Collection

Soil samples were collected from the Kellogg Biological Station Long-Term
Ecological Research Site (KBS LTER) located in Hickory Corners, MI. The dominant
soil series at the station are the Kalamazoo and Oshtemo series, which are fine-loamy and
coarse-loamy mesic Typic Hapludalfs, respectively. These soils have accumulated clay
and have a medium to high supply of plant nutrients and water with moist surface layers

that are slightly leached [28]. The three treatments focused on in this study were: an

51

agriculture site that receives amounts of fertilizer, herbicide, and conventional tillage
(CT) that are typical for the Midwest region and is on an annual corn (Zea mays L.) -
soybean (Glycine max L.) - wheat (T riticum aestivum L.) rotation; a historically tilled site
(HT) abandoned from agriculture in 1989; and a never-tilled site (NT) maintained with a
mid-successional plant community by annual mowing. A more detailed description of
treatments at the KBS LTER can be accessed on the World Wide Web at
http://lter.kbs.msu.edu/.

Initial studies on the presence of nirS and nirK in the CT and NT treatments were
conducted with soil collected in December 2003. Cores of 2.5 cm diameter and 10 cm
depth were collected from all five sampling sites within each treatment replicate. Soil
from the same treatment was pooled, homogenized by sieving (2 mm mesh), and stored at
4°C until DNA was extracted using the method of Zhou and colleagues [29].

Soil was collected again in December 2004 for an investigation of denitrifier
community composition and diversity. At the time of sampling, the CT treatment was
fallow and only wheat stubble remained. Soil was collected at depths of 0-7 cm and 13-
20 cm from two replicate plots of the CT, HT, and NT treatments. At each treatment
plot, a total of five cores of a 2.5 cm diameter were taken, the soil was divided into 0-7
cm and 13-20 cm portions, and portions from the same depth and replicate treatment
were pooled in a Whirl-Pak bag (Nasco, Modesto, CA). The soil was kept on ice until it
was brought back to the laboratory and stored at 4°C overnight. The following day, the
soil was sieved (2 mm mesh) to remove stones, plant biomass, and for homogenization.

Portions of soil were frozen in liquid nitrogen and stored at -80°C for later DNA

52

extraction. The remaining soil was stored at 4°C for pH and percent gravimetric soil
moisture determination.
Disturbance Experiment

In June 2005 nricroplots were constructed within the two previously sampled CT
replicate plots. In each replicate plot two 3 x 3 m nricroplots were set up side-by-side in
the Northeast corner. During the time of the experiment the main CT treatment had been
planted to corn. Prior to manipulation, all corn was removed from the microplots. One
microplot was a control in which no further disturbances were applied to the soil and the
other microplot was disturbed at intervals of once per week for a total of three weeks.
Weekly disturbances to the control microplot consisted of the following: rototilling the
entire microplot to a depth of 7 cm to simulate tillage and addition of water and fertilizer
to the center 1 m2 of the microplot. Water addition simulated an average rainfall of 0.5 in
per cm2 and fertilizer application was equivalent to 40 gal per acre with a 28% nitrogen
solution containing ammonium nitrate and urea. Each individual water and fertilizer
addition was a typical amount for the region. Five soil cores 2.5 cm in diameter and 7 cm
deep were taken hour the center 1 m2 of each plot before and after each round of
disturbances, one week after the final disturbance, and five weeks after the final
disturbance. The center 1 m2 of the plots was sampled to avoid edge effects from the
surrounding corn. Soil was sieved and processed in the same manner as the December
2004 samples.
Relative Disturbance Scale

In order to investigate the relationship between disturbance and proteobacterial

denitrifier diversity, a relative disturbance scale was set up for the denitrifier

53

communities in the December 2004 soil samples taken at 0-7 cm and 13-20 cm depths.
Disturbances were considered to be activities that could result in altered niche
opportunities for denitrifiers [30]. The scale was based on documented treatment
manipulations taking place at the KBS LTER including, tillage, herbicide/insecticide
application, fertilization, harvest/mowing, burning, and liming. Precipitation was also
considered a disturbance. Different treatments and depths were graded on a
presence/absence scale for each category of disturbance (Table 3.1).
pH and Percent Gravimetric Soil Moisture

To characterize physical aspects of the soils that have been shown to affect
denitrification, soil pH and percent gravimetric soil moisture were measured using
standard methods for LTER sites [31]. Soil pH was measured after 15 g of sieved soil
was mixed with 30 mL deionized water. The soil solution was allowed to equilibrate
with the atmosphere for a minimum of 30 minutes followed by pH measurement with a
pH meter (Orion Research Inc., Beverly, MA). Percent gravimetric soil moisture was
measured after 15 g of field moist soil was weighed and dried at approximately 50°C for
a minimum of 2 days. Soil moisture was calculated on a dry weight (dw) basis as 100 "'
((field moist weight - dry weight)/dry weight). Three replicate measurements of both pH
and percent gravimetric soil moisture were performed for each soil treatment and depth
(Table 3.2).
Carbon and Nitrogen Measurements

Total percent carbon and nitrogen of soil were determined by dry combustion.
Dried soil samples fiom December 2004 and August 2005 were ground using a mortar

and pestle. Three replicate 15 - 20 mg samples of soil were weighed out on a

54

.8888 88880 88.280 32.5. ”.5888 a .8: e

denies; u+.8...leggeeeieagasesmaeeeageaeeaseeageegaofi.

80:28.5.— 803 Eon—158a 883 v5 092:: deg—2.8.2 «e 25.95 32.805 :02? a E. E Eon—toes 8 we can 203 2c..— ouefi. e
83 8.5.82 u E o

82 a 2.28:8 58 882.58 83 8:: 28:82: a E ._

ca 8388.: 8:: 288858 u so .

 

 

 

 

 

8E 88 33 .55 33 .35 8.5 82: 2.2.5. 88880
- - e e n m N _ e 82.5. .585
2 a e n no on 2. a e 88m 8.8.85 88.3.
£3.62
+ + + 4+ + wage 8 083 3 805$
8:630 23:. 5 owes—o
+ 8382.. Ed:— MO 33on En
+ + 4+ + 2.. a 8885 use:
3.68:. «ea—a me
+ + -1 + 88.885 .858 0 e388 8228:
1 + 4+ + 2 use 2 ..2 egos a 8885 8:858..—
- - + + - 2 .8 oee 83cm 23,5 .8:
98:2.—
++ + + + 2.53% we 22.0.3 8 Sam 23 owe—E.
2:885 2 e5 o a 8885 23.88<
+ + ..\+ .\+ + + .37. 8883a E 38.—eon we 232$
2§ae§
.8 .85 =8 52.. z e5
i + - + - + 4+ + o as: 828.85 a 3.885 82.828.—
.8 we a... E as 8.2 .8 Z .8 8.2 .8 E .8 8.2 .8 2 seem 858.39
e858 12.805 1.2 12 .E 1.: 1.0 .5
88 883. 38 .8588 §_§

 

 

mega—5m am: we. a 3 8832.. 8 3 e288... B. 88:88— efixem "..n 3.2.

55

.38 00:08? .8 20.08:: 8%.:
_0>0_ mod": 05 :0 0:20am: 3:005:80 8: 0.3 .830. 088 05 .3 330:8 80G .0520: 98 3:08:00 8: 0.8 .80 8088580
:8 03:3 05 mo 86 05 800080.. 35885302 63783 88.: :83 0808080008 89: .880 83:50. :5 0m80>< 0

0000.820: 8 0.3 0.880 838% 50:80:00

88:8 203 008 .083. 05 38 03:80: a 028:8 30: .08. 0.0002200 0:8 0020:? 28808:... .8 pa 5:2 08a.
05 .3 330:0.“ 8.0a 008330.: 80880.: m .«o £000 .0 3:088:80:— 05 ”3:082:80:— 0 mo 03006 05 0: 3:80.: 2 8:: 38
28¢. 808280.. 80880: .0 mo :80 88.: 808080008 0:: “2:088:88 v .«o 0w80>0 05 m: @0830: mm 8:: 38 698002— .

 

 

< 2 0.8 2.0 m 80.8 :3. 80 3 ES 8
$882388 02 < 30.8 0; < 8: .e 8.0 so E. 880 8 008 .083.-
0 80.8 00.8 o 22.8 00.0 80 8-2 ...z
0 a 30 2: ... 80.8 8.: B 28.8 8.0 :5 E E
0 200. c 00.2 00 200.8 00.0 .8 8-2 ...:
0 90.8 8.0 0 :5. c 00.8 0 30.8 00.0 80 5.0 E
0 8.8 00.2 0 200.8 8.0 :5 8-2 8
0 5.8 8.0 0 80.8 00.2 028.8 00.0 80 3 B 38 00088:
:0 1382.828 . 2820: mom .2. . m: 08 088.00 02950 30:
._ 0.3.0 002

 

0.220 08:80 00 03E of 05 .2802 now .520: .:.: 8.0 202.

56

microbalance (Sartorius, Edgewood, NY) and packaged in tin capsules (Costech
Analytical Technologies, Valencia, CA). Combustion with a Costech Elemental
Combustion System was performed to calculate percent carbon, percent nitrogen, and
carbon/nitrogen ratio (Table 3.3). The standard for carbon and nitrogen content was
Phenacetin (Costech) and a combustion standard of Cyclohexanone—2,4 dinitrophenyl-
hydrazone (Costech) was also used.

Available carbon was measured from 1989 to 1996 as the short-term respiration
potential for soil collected from 0-25 cm. This was a routine measurement taken at the
KBS LTER, and as such the protocol and raw data are available on the KBS LTER
website. The average respiration potentials were: 115 2!: 5 pg C - g soil", 158 d: 5 pg C °
g soil", and 261 :1: 14 pg C - g soil'l for the CT, HT, and NT treatments respectively.
nirS Detection Limit

In order to obtain a general calculation of how much template DNA with nirS
would have to be present for detection with the PCR, various reactions were set up, each
containing a total of 50 ng DNA. Genomic DNA from Pseudomonas stutzeri JM300,
which has one copy of the nirS gene and a genome size of 4.03 Mb (ca. 2.3 x 105 copies
nirS/ng DNA) [32], was added to DNA extracted from the CT or NT treatments. The
primers used (nirS 1F and nirS6R) and PCR protocol followed were from Braker and
colleagues [25].
nirK Library Construction

Libraries were made from two replicate agricultural treatments and soil depths. In
one case, two libraries were made from the same treatment replicate and depth (HT 0-7

cm, Rep. 2). For each library, DNA was extracted from 0.25 to 1.0 g soil using an

57

52888 8888 83 38 882 88 38 .3888 a 882.8 38 .83 8.82.8 was. 8288
b58558 8: Pa .632 2:8 05 .3 330:8 Sam 88055.38 3 .55 8.3853 gees—2:308 m .«o 03.5.6 05 me cocoaom .

 

 

o 5.8 8.» o 5.8 m; m 83.8 3.2 a .88 .8 Z 85 .5
om $8 3.8 m< 9.8 8.8 m< 2 2.8 33 c .88 so 2 SS 5
m< $08 8.: m 88.8 25 m A38 «2 a .88 so 2 .888 so

< A38 8.: < 88.8 85 < 888 m3 2 .88 .8 3 3:8 .8 38 883..
a $18 2.8 a 898 8.8 a 28.8 who a .88 as 8.: ...z
Ba 58 8.: a 25.8 38 a 898 one 2 .88 .8 8.: E
o :88 3.2 0 A88 So a £8 8.... a .88 so 2 .2
8 $98 3.2 18.8 8.8 0 A88 :3 c .88 as g E
12.8 2.2 a 88.8 88 a 23.8 who a .88 .8 8.: E
a £88 8.8 a 88.8 8.8 a :98 :3 c .88 so 8.: E
8 A28 85 a 398 3.8 a :88 m: a .88 .8 2. .E
8.. 8:8 3.: n :88 :5 a 68.8 .84 2 .88 8° 3 E
8 828 8.2 a :58 85 8 A88 8.0 a .88 =8 8-2 B
8 3.8 3.8 a 28.8 8.8 8 A88 85 2 .88 as 8-2 .8
8 5.8 2.2 18.8 8.8 8 A88 who a .88 so 3 5

a $8 8:... 18.8 86 a £88 8.8 c .88 .8 E 5 88 83888

. zuu . 2.x. . 0.x. 38. 338m 388m 28

 

8% 38:8 .8 26 88 8882 .888 £380 .828 ”on 02:.

58

UltraClean Soil DNA Kit (MoBio, Carlsbad, CA). The primers FlaCu and R3Cu [33]
were used to amplify nirK genes. Amplification of DNA from soil was performed with‘l
U Taq polymerase (Invitrogen, Carlsbad, CA), a 1X concentration of the manufacturer’s
PCR buffer (20 mM Tris-HCl [pH 8.4], 50 mM KCl), 1.5 mM MgC12, 0.2 mM dNTP
mixture (Invitrogen), 0.01% Triton X-100, 0.02% BSA, 50 pmoles of each primer, and
10 to 50 ng of template DNA. PCR mixtures were incubated in a PTC-200 DNA Engine
gradient thermocycler (MJ Research, South San Francisco, CA) with the following
touchdown PCR protocol: (i) 3 min initial denaturation at 940°C; (ii) 10 cycles, with
each cycle lasting 30 s at 940°C, 40$ starting at 600°C and decreasing by 0.5°C each
cycle to end at 555°C, and 40 s at 720°C; (iii) 15 cycles, with each lasting 30 s at 940°C,
40 s at 570°C, and 40 s at 72.0°C; and (iv) 7 min at 720°C.

A low number of PCR cycles were used during library construction to reduce
PCR bias fi'om PCR drift. Twenty-five cycles were used as this was the minimum
number of cycles needed to produce enough PCR product to be visible when run on an
agarose gel and stained with ethidium bromide. Each nirK library was created using six
50 uL PCR reactions which were pooled and concentrated using a QIAquick PCR
Purification Kit (Qiagen, Valencia, CA). The PCR product was electrophoresed through
a 1.0% agarose gel and extracted from the gel with a QIAEX II kit (Qiagen) to eliminate
non-specific PCR products that might compete in the cloning reaction. Purified product
was cloned with a TOPO TA Cloning Kit for Sequencing with pCR 4 and One Shot®
TOPIO competent cells (Invitrogen) according to the manufacturer’s instructions.
Transformants were plated on LB agar containing 50 ug/mL kanamycin and clones

screened for the correct insert size using modified M13 primers F2 and R4" [34]. Prior

59

to sequencing, PCR product was cleaned up with ExoSAP-ITO (USB Corporation,
Cleveland, OH) to remove unused primers and dNTP’s. 1.3 uL of PCR product and 0.25
uL ExoSAP-TT® were incubated at 37°C for 30 min and the reaction inactivated at 80°C
for 15 min. Product was sequenced using a capillary sequencer (Applied Biosystems,
Foster City, CA) with dye-terminator fluorescent cycle sequencing technology at the
Michigan State University Research Technology Support Facility (MSU RTSF).
Diversity Data

NirK sequences of ca. 473 base pairs were aligned using the program ARB [35],
the primer sequence masked, and a distance matrix and phylogenetic tree based on
neighbor joining constructed that included all sequences. I-LIBSHUFF version 1.3 [36],
and in some cases the program UniFrac [37], was used to compare pairs of replicate nirK
libraries to determine whether libraries from replicate agricultural treatments and depths
were significantly different. Library comparisons in I-LIBSHUFF are performed so that
not only is Replicate Library X compared to Replicate Library Y, but Replicate Y is also
compared to Replicate X. This can result in cases where in one comparison the libraries
are significantly different, but in the other they are not. This indicates that one library is
a subset of the other and subsequently the libraries were considered to be not significantly
different.

DOTUR version 1.53 [38] was used to analyze the a-diversity of each library.
Input into DOTUR consisted of distance matrices based on neighbor joining from each
individual library which are used by the program to assign sequences to OTU’s at all
percent nucleotide similarities. DOTUR output was used to calculate Simpson’s diversity

indices (-lnD), Simpson’s evenness measure ([l/D]/S, where S is equal to the number of

60

OTU’s present in each library rarified to the number of clones in the smallest library),
Good’s coverage ([1-(n/N)]*100, where n equals the number of singletons and N equals
the total number of sequences), and rarefaction curves at three nucleotide similarity
cutoffs of 94%, 97%, and 99%. Output fiom DOTUR was also used to calculate
rank/abundance curves in which OTU’s fiom libraries were plotted from most to least
abundant versus their relative abundance.

In addition, B-diversity was determined also using both DOTUR and
I-LIBSHUFF. DOTUR was used to separate sequences fiom all libraries into OTU’s at
94%, 97%, and 99% nucleotide similarities. A program was written (GeneMatrix) by
personnel at the Ribosomal Database Project (RDP) to convert the DOTUR .list file into
a matrix that could be used as input in the program Estimates version 7.00 [39]. Jaccard
dissimilarity matrices created with Estimates data were converted to dendrograms using
the program MEGA version 3.1 [40].
rrn Copy Number & Growth Rate

Dissirnilatory nitrate reducing microorganisms with known ribosomal RNA (rm)
operon copy numbers were chosen to investigate whether there was a relationship
between rm operon copy number and growth rate. Nine organisms were used in the
study, and of these, the growth rate of six (Escherichia coli K12, Pseudomonas
fluorescens ATCC 33512, Pseudomonas stutzeri JM300, Ralstonia metallidurans,
Serratia marcescens, and Shewanella putrefaciens ATCC 8071) were determined in the
laboratory. The growth rates of the remaining three organisms (Chromobacterium
violaceum CS-l , Magnetospirillum magnetotacticum MS—l, and Rhodopseudomonas

palustris BKl) were obtained from the literature.

61

Strains were grown on Tryptic Soy Broth (TSB) with 5 mM KNO3 that was
prepared anaerobically (Helium headspace) for growth of organisms capable of
denitrification or dissimilatory nitrate reduction to ammonia (DNRA). Growth rates were
obtained by monitoring the ODwo of replicate cultures (11 = 3 to 6) during incubation at
25°C and 180 rpm. Doubling times ((1) were calculated during the exponential phase of
growth and maximum growth rate (mm) calculated as ln(2)/d. The impact of phylogeny
on the results was assessed using the computer program CONTINUOUS (http://
www.mbic.rdg.ac.uk/meade/Mark/).

Nucleotide Sequence Accession Numbers

Partial nirK gene sequences were deposited in the EMBL, GenBank, and DDBJ
sequence databases under accession numbers DQ782971 - DQ783217, DQ783219 -
DQ783225, DQ783227 - DQ783279, DQ783281 - DQ783813, DQ783815 - DQ783894,

and DQ783896 - DQ784090.

RESULTS:
Percent Nucleotide Similarity Cutoffs

Nucleotide similarity cutoffs of 94%, 97%, and 99% were used to assess whether
nirK sequences belonged to the same denitrifier species. A range of cutoffs was used to
confirm that the trends found did not depend on a particular cutoff. 94% was chosen as
the lowest cutoff due to the fact that in a recent survey of conserved genes from 70 fully
sequenced genomes, an average nucleotide identity of ca. 94% corresponded to the
traditional 70% DNA-DNA reassociation standard currently used to define microbial

species [41]. Higher percentage cutoffs were used as some microbial species (e.g., E.

62

colr’) have been shown to have ca. 99% nucleotide similarity between strains analyzed
with multi-locus sequence tagging (MLST). Additionally, higher percent cutoffs may be
needed since there is evidence for lateral gene transfer of nirK between species. In order
to test whether lateral gene transfer would cause an underestimation of the number of
species, 40 nirK sequences from 36 different species of known denitrifiers were analyzed
with the computer program DOTUR which assigns sequences to OTU’s [38] (data not
shown). At 94%, 97%, and 99% nucleotide similarity cutoffs, the number of species was
underestimated indicating that if lateral gene transfer were to occur amongst denitrifiers
in soil, the number of denitrifier species predicted using nirK may be an underestimate of
the true number of species.
Detection Limit of nirS

Using two different sets of nirS primers and protocols [25, 33], little to no
detection of the correctly sized product was obtained in PCRs with environmental DNA
from the CT and NT treatments. Reactions with CT DNA consistently resulted in a very
low amount of product being produced, while nirS was never detected in reactions with
NT DNA. To investigate, a series of PCRs were performed with DNA fi'om P. stutzeri
JM300, which possesses the nirS gene. Genomic DNA from P. stutzeri was added in
various amounts to CT and NT treatment DNA. As a low amount of PCR product was
found in CT reactions with no added P. stutzeri DNA, the lower limit of detection was
determined to be 1 pg of P. stutzeri DNA as this amount was the lowest causing an
increase in band intensity on an agarose gel. The lowest amount of P. stutzeri DNA
resulting in a visible band when mixed with NT DNA was 4 pg (Fig. 3.1). Therefore the

primers should be capable of amplifying nirS genes even if DNA known to contain the

63

NT + R stutzeri DNA CT + R stutzeri DNA

100 bp ladder

0 0| ng
0.05 ng
0.1 ng

   

0.05 ng
0 l ng
0.1 pg

GOODODODOD
manna.
~NV\OOO

 

0.1 pg
3 0.01 ng

0000 00
0.0- O.
v-‘N OO

.-

1000 hp 1000 bp
800 bp 800 bp
600 bp

600bp

l f It!!!"

‘ I m. ‘3;- 100 bp ladder

 

Figure 3.1: Agarose gel of PCRs of genomic DNA from nirS-bearing Pseudomonas
stutzerz' JM300 mixed with environmental DNA from either the NT or CT treatments.
Picogram and nanograrn amounts of DNA listed above lanes refer to the amount of P.
stutzeri genomic DNA present in a PCR containing a total of 50 ng DNA.

proper template made up only 0.002% to 0.008% of the total DNA in the reaction. Due
to the fact that little to no nirS product was obtained from environmental DNA, the
denitrifier population possessing nirK was focused on.

Analysis of nirK Replicate Libraries

I—LIBSHUFF was used to determine whether pairs of replicate nirK libraries were
significantly different from each other. Initially, eight pairs of libraries were compared
and in six cases the libraries were not significantly different, while in two they were
(Table 3.4). Libraries from treatment replicates l and 2 were significantly different from
HT 0-7 cm and NT 13-20 cm and further analysis was performed to determine the cause
of the differences.

A second library fi'om HT 0-7 cm replicate 2 was constructed fi'om a new DNA
extraction to confirm that the differences in the replicate l and 2 libraries was not the
result of bias. The two libraries from HT treatment replicate 2 (library replicates 2a and
2b) were not significantly different from each other when subjected to I-LIBSHUFF
analysis. In addition, a P test was performed using the program UniFrac. Both the
I-LIBSHUFF and P test analysis results were in agreement, indicating that HT treatment
replicate l and 2 communities do differ at a depth of 0-7 cm. It was noted that one OTU
dominated in the replicate 2 libraries that was not present in the replicate 1 library. After
this OTU was removed and the I-LIBSHUFF analysis repeated, the replicate l and 2
libraries were not significantly different (data not shown). Therefore, the differences in
the HT 0-7 cm libraries of replicate 1 and 2 were due to the presence of one prominent

OTU in the replicate 2 community.

65

88888 82 u <2

8888 88888 n 8 88.88 82 u 92

8.8 u a 8 8888 288888 n .

8888-18 888888 88:88 8... 8 8888 888: 888 28.8 8888880

.N 08022—02 E08802 80¢ 0808 803 85.5: 9.3 23 _ 38:80.. E28008 89¢ 0808 was 80 We E 89¢ En: 0:0 .

 

 

 

~88 28.8 228888 80.8 ~88
22 8 888 am 8 888... 22 8x 85 am e _ 88H 22 .8 888 .882
02 _ 80M .3:an . 82 _ @3—
<2 <2 <2 <2 <2 28.8 <2 <2 8 98 88
<2 <2 <2 <2 <2 82.8 <2 <2 8 88 88
<2 <2 <2 <2 <2 .. 8.8.8 <2 <2 298 88
<2 <2 <2 <2 <2 .. 28.8 <2 <2 8 2 2 88
<2 <2 <2 <2 <2 ._ 88.8 <2 <2 238 88
<2 <2 <2 <2 <2 188.8 <2 <2 8 2 2 88
8.8.8 . 83.8 . 88.8 8:3 888 <2 23.8 ._ 88.8 2 a N 8m
828 R88 .. 88.8 .. 88.8 :83 <2 28.8 82.8 N a _ 8m
:8 we :8 E. =8 8-2 =8 E :8 8-2 :8 we :8 8-2 =8 E 8888880
888888 8880 E 2 E ...-2 .8 8

888885 88%

 

88.2 .88:

 

888: 02.8 888:8... 8 888.2 8528i :88 8:83— 3 289

The NT 13-20 cm replicate libraries were also analyzed with a P test that
confirmed the significant difference between them. There were only a few OTU’s that
were unique to each library and none dominated the library. It was concluded that the
libraries reflected a true difference in the communities of both replicates.

Denitrifier Community Analysis

Differences between communities, or B~diversity, were compared by using
I—LIBSHUFF and cluster analysis with Jaccard dissimilarity matrices on nirK libraries
from soil collected in December 2004 and August 2005. The I-LIBSHUFF analysis
showed that there are similarities between the communities of the CT libraries at 0-7 cm
and 13-20 cm libraries from December 2004 soil and the CT 0—7 cm libraries from
August 2005 soil (Fig. 3.2). The 13-20 cm library from the HT treatment was similar to
the CT 13-20 cm library, but not to the CT 0-7 cm library. Only one replicate library
from HT 0-7 cm showed a similarity to other libraries.

Jaccard dissimilarity matrices were used to create dendrograms of the
relationships between treatments based on OTU composition. Unlike I—LIBSHUFF, the
Jaccard measure does not take the abundance of OTU’s into account and is therefore
purely a measure of the differences between community compositions. At all three
nucleotide similarity cutoffs, libraries grouped into four clusters (Fig. 3.3). Cluster I
consists of 0-7 cm libraries from the CT treatment from both the December and August
timepoints. The 13-20 cm libraries from both the CT and HT treatments form Cluster 1].
The HT 0-7 cm libraries form Cluster III and the 0-7 cm and 13-20 cm libraries from the

NT treatment comprise Cluster IV. These results corroborate those found in the

67

.8050 :8 80¢ 80.80.¢=8 b88880¢¢8w_m
0.803 8088888: 85 88068 ...-8003 8003808. 8398.888 02 .08 8088088 2 8885 808888088 05 8 ENE: 88 -¢0 803888
88 m8 808800.88 0880 80¢ 8.88.58— 0 8885 30.8.88 803088-080 a 280 8080.¢¢8 288888088888 80: 0.803 808.889: 038 805
0888068 88003 8003800 8398.88 @0880: 03,—. 280288800 "E Emma-1 w 80 w .¢0 308 8 020288 8800 8:808 8888
28808800.: 88003809 288888800 0888w¢ 05 8 8003 880-00 3 8008080808 82-8808— ¢0 83888 08: :0 98688088080
6.80.6808:- .8858800 82-058— 038 020288 805.58— 05 880038088 magma-H A 8 88.588800 80.80-¢=0
28008-28888 0.803 8885 880880.88 088808—8808 89¢ 80888.58— 2: 8080.808 9.003 2.08m 80.80.8888 b8880¢¢8w8m
8088 0.803 :02? 888088 :8 88888 80888808 888-8388808888 8.080 89¢ 80888.5: 0880:8808 038 05 8080.808 88003 033
.888888 8888888881 88 888828 888888 888383 8888888888 288 8 888888282 88.8.8 < 8.8 288.8

Ndlmgm wwfizgm Nae—30y.

Eu ONU- m. 80 h- o

 

 

. ans
N ...Mv — mun—OZ N Q0.“ N MW _ mQOM— n N a — wavy—
EO F: O

FCU \llc EU “IO EU “I o
...z :8 8 H 8888 5

_
n L. _|l_

we: 80:88.30: mccm 8888883.

 

 
 

 

68

 

CT Dist. 0-7cm (Rep 2)
CT 0-7cm (Rep 1)

CT Cont. 0-7cm (Rep 2)
CT Dist. O-7cm (Rep 1)
CT Cont. O-7cm (Rep 1)
CT 0-7cm (Rep 2)
CT l3-200m (Rep I)
HT l3-20cm (Rep 1)
CT l3-200m (Rep 2)
HT l3-20cm (Rep 2)
HT 0-7cm (Rep 2a)
HT 0-7cm (Rep 1) Cluster 111
HT 0-7cm (Rep 2b)
NT l3-20cm (Rep 2)
NT l3-20cm (Rep 1)
NT O-7cm (Rep 1)
NT 0-7cm (Rep 2)

 

 

Cluster 1

 

 

 

 

 

 

Cluster [1

 

 

 

 

 

 

 

_* 8L rL—L 4:181

 

 

Cluster IV

 

 

l 1
U U

40% 30% 20% l 0% 0%

Figure 3.3: Dendrogram based on a Jaccard dissimilarity matrix at a 94% nucleotide
similarity cutoff. Dendrograms from the 97% and 99% nucleotide similarity cutoff are
not shown as the same clusters were formed. Scale represents percent dissimilarity.

69

l-LIBSHUFF analysis and also show that the relationships seen between communities are
due to differences in community composition and not just abundance of OTU’s.
Diversity, Richness and Evenness

Because differences were seen in community composition, the diversity of the
communities was compared. To assess whether differences existed, the overall diversity,
richness, and evenness of the nirK libraries was calculated at three nucleotide similarity
cutoffs (Tables 3.5, 3.6, and 3.7). Simpson’s Diversity index was used as an overall
diversity index as it incorporates the two aspects of diversity, richness and evenness.
Richness and evenness were also calculated separately. Since the number of clones in
each library differed, richness was calculated as the number of OTU’s observed when the
libraries were rarified to the size of the smallest library. Simpson’s evenness was
calculated from the Simpson’s Diversity index and the rarified number of observed
OTU’s. Good’s coverage of each library was compared in an effort to determine whether
differences in sampling effort between libraries existed. Coverage varied from 58 to
86%, 53 to 82%, and 37 to 69% at the 94%, 97%, and 99% nucleotide similarity cutoffs,
respectively. A 2-way AN OVA assessed whether any differences seen in diversity,
richness or evenness were due to agricultural treatment, soil depth, or an interaction
between the two. At all nucleotide similarity cutoffs there was no significant effect found
for Simpson’s Diversity index, ratified number of OTU’s or Simpson’s evenness due to
either depth or a depth "' agricultural treatment interaction (P > 0.05). However, at the
94% nucleotide similarity cutoff, there was a significant effect from agricultural
treatment (Fig 3.4) on both the ratified number of OTU’s (P = 0.0447) and Simpson’s

evenness (P = 0.0476). Pairwise t-tests indicated that it was the richness and evenness

70

 

88.88 cm 88 8 3 8m 88088 88 8.80 80283880 5

 

8888.80 8.8 82 888 888. 88 88388 88 8.88 8.883880 8
8.88 8m 83 8.0 8 8m .8388 88 28 888850 8
8.88 mm 88.8 R R 88 88388 88 8.88 88880 S 88 888838
8.8.88 R 8.3 08 8 88 88388 80 883.8 .:8
2.88 888 :8 a 8 88 .8388 888.888 8.2
8828 E 3.88 888 «0 8m .8388 88 8.88 :8
3.88 R 8.8 8888 8 88 .8388 88 8.88 8.2
8.88 8 88 Q a 8a 88388 so 8838 .8
80.80 R 8.8 8 8.88 88 88388 85 888-888 :8
8.88 mm 38 R 888 8888 880888 88 8.88 E
02 «N 88 8: a 888m 88388 85 8.88 8.88
$.88 88 ~08. E 08 88 88388 88 28 8
888.88 8 «8.88 8:. m8 8N 88388 88 8-88 5
$.88 8.8 N; 888 8 88 88388 88 8-888 S
88 NM 3.». 8.0 8.8 88 .8388 88 E8 8 8.88888
$8 2 8.8. S 8 88 88388 88 E8 .8 82880888
89.8 8.88.5 8988.8 0888050 88880 2888.838 0 88888088 88888828 3880 293
00088880>m .80 80888882 8.8088888 0.88000 .80 80888882
8.8888888 38.8858 085,888 8.888888%

 

.8825 88888888880. 0888888282 :3 a s 8888888 8888.. .88. 008580 8.8 88888830 8.8.6288 8?. 0888888.

71

 

 

88.0 mm 08.8 80 88 88 88388 88 8.0 03808880 80

08.0 mm 00.8 888 888 88 88388 .8 8-0 02883880 80

08.8 8.8 88.8 00 0... 88 88388 88 8-0 888000 .80

00.0 08 08.8 00 mm 88 88388 85 8.0 80888800 .6 0008 8300.88
80.0 mm 88.8 08 8 88 88388 88 08-08 8.2

00.0 0m 0m...- 88 80 88 88388 88 08-88 8.2

00.0 88 8.8.8 08 80 88 88388 a... 8-0 8.2

80.0 08 08.8 88 00 88 88388 88 8-0 8.2

08.8 08 80.8 88 0m 88 88388 88 08-08 .888

80.0 mm 00.88 .80 :8 88 880888 88 08-88 .888 2

7

80.0 888 2.8 88 888 8888 88388 a... 8-0 8.88

80.0 8.8. 00.8 “8 08 808 880888 88 8-0 8.88

08.8 8.8 88.8 00 08 88 88388 88 8-0 8.88

8.0.0 88. 88..8.. 80 888 88 00888 85 08-88 .6

8.8.8 88. 88.8 mm 00 88 00888 88 08-88 80

08.8 mm mom 8 8m 88 880888 88 8-0 .80 8008
8.0.8 08. 08.8. mm 00 88 880888 85 8-0 .6 82880888
80885 0.3.8.0 588.8 03260 0088080 080088880188 0% 88888008 808802.8- 35 038%

00088086 .80 80888882 880888 0.0000 .80 80888882
8.0808080 88.8858 888028888 8.080580

 

8.80800 8888088880 0888888002 .880 a 80 8088088888 888.5 80 0888260 05 8880-88080 8888828888 ”0.88 0880.8

 

00.0 00 00.0 00 00 80 0380 88 0.0 805000 .00

 

00.8 8.0 00.0 00 00 88 0380 85 0.0 80285005 .00
00.0 08. 80.8. 08. 00 80 0380 85 0.0 888800 00
08.8 00 00.0 00 00 88 0380 85 0.0 800850 00 0000 0803..
00.8 00 00.0 00 00 80 030 :80 00-08 .02
3.8 00 00.0 00 00 88 88380 88 00-08 .02
80.8 00 00.0 00 00 80 0380 88 0-0 02
88.8 00 00.0 00 00 88 88380 88 0-0 02
00.8 00 00.0 00 00 80 0380 88 00-08 E
00.0 00 00.0 80 8.0 88 0380 85 00-08 E
00.0 00 08.0 00 80 800 0380 88 0-0 .5
00.0 00 00.0 80 00 800 0380 88 0-0 008
00.8 00 00.0 00 00 88 030 80 0-0 0:
80.8 00 00.0 00 00 80 0380 88 00-08 00
00.8 00 v0.8. 080 00 88 0380 88 00-08 00
00.8 00 00.0 00 00 80 88380 =5 0-0 00 8.000
00.0 08. 00.0 00 00 88 0380 88 0-0 .00 308080
38.-“c 0.3.8.0 803 own-82,00 0080080 880020038 0% 8.80qu 4003.8. 20a 2880me
00008080>m .00 000888 Z 0805 00080-8905 0.0090 .00 0008052
0.8880000 02-0858 0.0000850

 

.0300 08380880 0008882 .000 0 00 8:80: 080: 00 008260 000 800030 0003388 H0.0 08000

73

1 —0— 0-7 cm
36- + 13-20 cm

28.

26-

Rarified Number of OTU's

241

 

 

I

Conventional Mid HT (HT) Mid NT (NT)
Agriculture (CT)

B. L8“ —0— 0-7 cm

l6“ +13-20 cm

L4-

[.2-
1.0-
0.8-

0.6-

Simpson's Evenness ((l/D)/S)

0.40

0.2-

 

 

] j
Conventional Mid HT (HT) Mid NT (NT)
Agriculture (CT)
Agricultural Treatment

Figure 3.4: Results of a 2-way ANOVA using data from a 94% nucleotide similarity
cutoff. Significant effects due to agricultural treatment were found for both richness and
evenness at a = 0.05. Graphs show averaged data from 2 replicate libraries from the
same treatment and depth. A) Richness assessed as the number of OTU’s rarified to the
number of clones in the smallest nirK library. B) Simpson’s Evenness. Error bars are
standard errors.

74

measure values from the CT treatment at 0-7 cm that were the most different from those
of the other treatments and soil depths.

Another method used to compare species abundance data is to construct
rank/abundance curves [42]. These plots highlight communities that differ greatly in
evenness by showing whether the community is dominated by one or a few OTU’s.
Rank/abundance curves were constructed from data at the 94% nucleotide similarity
cutoff, with those replicate libraries that were found to not differ significantly in the
I-LIBSHUFF analysis represented by an averaged curve (Fig. 3.5). The CT 0-7 cm
treatment libraries had the flattest curve indicating few, if any dominating groups. All
other libraries differed in this regard, some significantly as shown with a Kolmogorov-
Smirnov two sample test. Both the HT 0-7 cm replicate 2 and NT 13-20 cm replicate 2
rank/abundance curves differed from the CT 0-7 cm libraries (P < 0.05 and 0.05 < P <
0.10, respectively).

In order to determine whether the same nirK OTU’s were dominant in the NT and
HT clone libraries, the NT OTU’s were ranked according to the percentage of the library
they comprised and compared with the same OTU’s in the HT and CT libraries. The 0-7
cm libraries were concentrated on as this zone of the soil contains the majority of
microbes and is the location of most metabolic activity. There were six NT 0-7 cm
OTU’s (out of a total of 47) that accounted for approximately 50% of the total sequences
(Table 3.8). These same OTU’s comprised ca. 25% of the HT 0-7 cm clone libraries, but
only ca. 1% of the CT 0-7 cm libraries. This has implications for determining which
denitrifiers in NT and HT might be responsible for contributing to the low N20 fluxes

found in these treatments.

75

A. 0.30
+CT 0-7 cm (Reps l & 2)
I -0- HT 0-7 cm (Rep 1)

 

 

 

 

 

 

025 + HT 07 cm (Rep 2a & 2b)
3 . +NT 0-7 cm (Reps 1 & 2)
S
'6
=
3
.0
4:
G)
>
'5
£3
0)
a:

0 10 20 30 40 50 60
B 0.30 -
- +CT 13-20 cm (Reps l & 2)
O 25 q +HT 13-20 cm (Reps l & 2)
' W "4" NT l3-20 cm (Rep 1)

8 + NT 13-20 cm (Rep 2)
5 0.20 -
'5
g 1
.0 0.15
<
3
'0 0.10
a
7-
“ 0.05

0.00

 

 

Rank

Figure 3.5: Rank/abundance curves of nirK libraries at a 94% nucleotide similarity
cutoff. Treatment replicates that were not significantly different in a I-LIBSHUFF
analysis were combined to obtain an average rank/abundance curve. Treatment replicates
that were significantly different were kept separate. A) Libraries from a depth of 0-7 cm.
B) Libraries from a depth of 13-20 cm.

76

Table 3.8: Dominant nirK OTU’s from NT 0-7 cm and their Percent Abundance in HT
and CT at 0-7 cm "

 

 

 

 

Percentage of Library (%)
NT OTU Rank NT HT CT
1 11.8 0
2 9.2 7.9 0.9
3 7.9 1.4 0
4 7.9 O 0
5 5.3 0.5 0
6 4.6 13.9 0
Total 46.7 23.7 0.9

‘ OTU’s that represented 2 4% of the combined NT 0-7 cm clone libraries were
considered dominant OTU’s

77

Relationship Between Diversity and Productivity

Productivity was measured as the total percent carbon (%C) of the soil
environment that each library was constructed from (Table 3.3). There was no difference
in %C (or %N) between the 0-7 cm and 13-20 cm depths of the CT treatment, the 13-20
cm depths of the HT treatment, and the 13-20 cm depth of the NT treatment. The 0-7 cm
HT and NT treatments were significantly different from all other treatments and depths in
%C. Linear regression was performed to determine whether there was a relationship
between %C and overall diversity, richness, or evenness. No significant relationship was
noted at any of the percent nucleotide cutoffs and with all three diversity statistics (all P >
0.05).
Relationship Between Diversity and Disturbance

In order to investigate whether there was a relationship between diversity and
disturbance, each agricultural treatment and soil depth was rated as to its level of
disturbance (Table 3.1). Disturbances were considered to be events that could affect
denitrifier niche opportunities [30]. At all percent nucleotide similarity cutoffs an
increasing trend in Simpson’s Diversity was noted with increasing disturbance (Fig. 3.6).
To investigate whether this relationship was significant, Spearman’s Rank Correlation
Coefficient was calculated. At the 94% and 99% nucleotide cutoffs there were weakly
significant, positive relationships between disturbance and Simpson’s Diversity index,
the rarified number of OTU’s, and Simpson’s evenness (Table 3.9).

To determine whether the denitrifier diversity and disturbance relationship would
follow the predictions of the IDH a disturbance experiment was conducted in the summer

of 2005. It was hypothesized that an increase in disturbance level at the most highly

78

_ I 94%
4.8 O 97%

' A 990/
4.4 - °

4.0 - i
3.6 4 } §

i

3.2-

§
9
2.8! }

2.4 -

I'I'IF—o—fl-H

Simpson's Diversity Index (-lnD)

 

 

I I I I I I

2 3 4 5 6 7

Relative Disturbance
(0 = Low Disturbance, 7 = High Disturbance)

c.

Figure 3.6: Relationship between Simpson’s Diversity Index and relative disturbance at
94%, 97% and 99% nucleotide similarity cutoffs. Data is from libraries created from soil
collected in December 2004.

79

Table 3.9: Spearman’s Rank Correlation Coefficient for Diversity Measures (December
2004 nirK Libraries)

Spearman’s Rank Correlation Coefficient (p) '

 

 

% Nucleotide
Similarity Simpson’s Diversity Rarified Number Simpson’s Evenness
ClltOff (4111)) Of OTU’S (S) (E1 /1))
94% 0.714 ° 0.600 c 0.671 c
97% 0.486 0.529 0.529
99% 0.643 ° 0.771 b 0.600 °

 

' Spearman’s Rank Correlation Coefficients were calculated based on average diversity
measure values for replicate soil treatments and depths.

b A significant correlation at a = 0.10 (N = 6, df = 4).

c A significant correlation at a = 0.20 (N = 6, dt = 4).

80

disturbed site would cause a decrease in diversity. NirK libraries were created from DNA
collected fiom both control and disturbed microplots located within the CT treatment. A
paired, one tailed t-test indicated that there was no significant difference in overall
diversity, richness, or evenness between the control and disturbed plots (P > 0.25).
rrn Operon Copy Number and Growth Rate

Previous work has shown evidence for a trade-off occurring between rrn operon
copy number and microorganism growth rate. To determine whether a relationship
existed specifically for dissimilatory nitrate reducers, denitrifiers and organisms capable
of dissimilatory nitrate reduction to ammonium (DNRA) with known rrn operon copy
numbers were identified. Growth rates under dissimilatory nitrate reducing conditions
were obtained either from the literature or were determined in the laboratory (Table
3.10). A significant, positive relationship (R2 = 0.8115, p = 0.0009) was found between
rrn operon copy number and growth rate (Fig. 3.7). Since species cannot be considered
independent units due to their phylogenetic relationships, the impact of phylogeny was
investigated [43, 44]. The computer program “Continuous” was used in the investigation
to show that phylogeny was independent of the relationship between growth rate and rm

copy number (9. = 0).

DISCUSSION:

The primary goal of this study was to compare proteobacterial denitrifier
community composition between sites differing in N20 flux, which is an important
ecosystem function, and history of agriculture. Analysis showed that the CT treatment

communities did not differ fi'om each other at 0-7 cm and 13-20 cm. This is most likely

81

2380.808: 05 80.8.0 005300 0008 538m .800 0580.0 0088 0803 0.80-8.80 80.30835 .8

830.30.33308055 380003

MOE. 05 880 00008080 0000880038 8.0300082 0200088208800 05 80 3E Wm 008—3— 0085 05 wfims 00880.0 0.80? 080088888 8300 88.: .

 

 

08030 0.88: 80.0 .8. 00.0 0 588808 088088002200 8000 000< 30.880000300502000
PE 38.0 m SEE-.5 00800008000. Two 88500303 53.800000008800005
.8030 0::- mcd 0n mmd 0. (JED 00080000080008-» 0.200.000.8880 0.00-0.00m.
888080 008E. 80.0 um 30.0 0. «£75 0500000080808-» SM .8000 030.2000um
3080 0888.8- 806 “m vmd m 8080588809 08000000008080-» N 8 mmm 00.30 020000003§§0808§000~
88030 02,—- ocd H mod 0 008-0080008008 08.808000008009n 0.0000088800000008 0.808030%
.8080 008E- Sd « mad 80 005-00809 00000000000088-» 00$):- 0000300 0.080020088008000
GE mod N Sambaoa 0808030080800 55 0.800200% 0.3050838030880050
FE 3.0 m 8050888008 08.808080000800908 TmE 50800000000008302 ES~2300~02M§<
88820-838 .8 080 0 8.00 . 888078 0000 58288 $80 808880080
838 530-00 838 53000 08000 8.8 80 80.8-0.588809 omuocoworfim

 

008 530% 08.80 000800888 8300 0808000 0.8.8 0003000 030008828 05 058.8800 08 00088 08080-8030 88:.m 050.8.

82

2“
O
1

R2=0.8115, p=0.0009

Q
00
l

A

C
O\
a

Growth Rate (hr")
:3
A

 

 

 

rrn Copy Number

Figure 3.7: Relationship between dissimilatory nitrate reducer growth rate and rrn copy
number under anaerobic conditions.

83

due to tillage, which homogenizes soil, making it a likely cause for the similarities seen
between the upper and lower soil communities. The CT 13-20 cm and HT 13-20 cm
communities were very similar, however the HT 13-20 cm community differed from that
found in CT at 0-7 cm. This is evidence that by December, enough time had passed for
the CT community to begin differentiating based on soil depth. In addition, there is
evidence for changes in the HT treatment denitrifier community occurring more quickly
in the upper portion of soil as compared to the lower portion. At the time of sampling it
had been 15 years since HT was used for agriculture. l-LIBSHUFF analysis shows that
the community in the upper portion of soil bears few similarities to CT, but is also not
similar to the NT treatment. Since the community in the lower portion of soil still shares
similarity with the lower portion of CT, it seems that community change is taking place at
a faster rate in the top 7 cm of soil. Plants are known to influence microbial communities
in soil, most likely through carbon inputs [49-53]. Since %C and %N are significantly
higher at 0-7 cm than 13-20 cm, selective pressures probably differ at each depth
resulting in different communities of denitrifiers. The fact that the HT community still
shows a detectable impact from agriculture after 15 years is not surprising. Previous
work at the KBS LTER has shown that there is a long-lasting effect on microbial
communities after soil is used for agriculture [54, 55]. This is most likely due to the fact
that soil carbon and nitrogen can take greater than 60 years to recover to pre-agricultural
concentrations and qualities [56, 57].

Community composition and diversity are different concepts in the sense that it is

possible for communities that differ in composition to have the same diversity. In

84

addition, there are many ecological theories regarding the maintenance of diversity, and
for these reasons overall diversity and factors that might influence it were investigated.

Differences in diversity measures were not related to the productivity of the
environment. This is not surprising as previous studies indicate that relationships
between diversity and productivity can vary based on scale, trophic level, or directness of
the productivity measure [21, 22]. The use of total %C as a measure of belowground
productivity for denitrifiers does not account for the availability or quality of carbon and
may not have been a direct enough measure to capture relationships between denitrifier
diversity and productivity.

There was a weak, positive relationship between disturbance level and overall
diversity, richness and evenness. Subsequent increases in level of disturbance did not
result in a significant change in overall diversity, richness, or evenness indicating that the
highest disturbance level employed was not high enough to cause increased denitrifier
mortality. The lack of a significant change may have been due to the lack of carbon
inputs as all plants were removed from the microplots. While the results of these
experiments do not support or refute the predictions of the Intermediate Disturbance
Hypothesis (IDH), increasing disturbance through agricultural management is related to
increases in denitrifier diversity. Supporting this is the fact that there was a significant
difference found in richness and evenness based on agricultural treatment. Evenness in
particular plays an important role in the difference seen between communities.
Dominance of particular OTU’s in the HT and NT treatments that are not present or at
low number in the CT treatment may be important in explaining differences in N20 flux

seen in these treatments.

85

The increase in diversity measures with increased agricultural management seen
in this study confirms previous results obtained in a study in which denitrifier diversity
was measured in 0-25 cm cores fiom the CT and NT treatments. NosZ gene diversity
was measured using RFLP with the result that the CT treatment had higher overall
diversity, richness and evenness [58]. DNA was extracted from soil at about the same
time of year as in this study, but took place in 2001. The consistency in trends seen over
the course of three years shows that the diversity measured is a consistent, and not
transient, property of these denitrifier populations.

Most studies exploring differences between bacterial communities rely on
comparisons between only one replicate of each treatment. To determine the potential
impact of this practice, an investigation into whether replicates of agricultural treatments
have the same diversity and harbor the same denitrifier communities was performed. In
the majority of paired libraries, the communities were the same. However, in two cases
there were significant differences in the communities as reflected in the libraries created
from them. The different libraries were from HT and NT treatments which have much
higher plant diversity than the CT treatment. One possible explanation for the
heterogeneity found within replicates is that it is due to differences in local plant life.
However, if this were the case, it would be expected that the 0-7 cm libraries from the NT
treatments be significantly different as well since the majority of plant roots are found in
the 0-7 cm depth of the treatment. As this was not the case, the difference noted in the
NT 13-20 cm libraries is not readily explainable. It may be that the soil properties of the
NT treatment are more heterogeneous at a lower soil depth causing microbial populations

to differ. The differences seen within treatments is consistent with past work in which

86

the variability of 16S rDNA TRFLP profiles fi'om the total microbial community of the
CT, HT, and NT treatments were compared [54]. The CT community was less variable
than the HT and NT communities, significantly so in the case of the NT community.
Regardless of the underlying causes of differences within treatment replicates, this study
demonstrates the need for replication in sequence libraries when molecular surveys of
soil communities are performed.

In addition to investigating the differences between communities in treatment
replicates, the occurrence and extent of PCR bias was investigated. In any study of
diversity based on molecular methods, questions regarding bias must be addressed. PCR
bias has been observed to occur due both to PCR drift and PCR selection [59]. In order
to minimize the contribution of PCR drift, which is due to random variation in the early
cycles of PCR, the minimum number of PCR cycles needed to produce a visible band on
an agarose gel was employed and multiple PCR reactions fiom a particular treatment and
depth were combined before cloning occurred. If PCR selection were to occur, this
would result in preferential amplification of certain nirK OTU’s so that their observed
number in clone libraries would not reflect their abundance in the soil environment. This
preferential amplification should occur in a repeatable manner so that libraries would not
differ significantly in the frequency of these OTU’s. To address this issue, the libraries
fiom HT 0-7 cm (Replicates 2a and 2b) and NT 0-7 cm (Replicates 1 and 2) were
compared. The three OTU’s that were found in the highest frequency in both treatments
were compared with a x2 test to determine whether they were found in a 1:1 relationship
as would be expected if PCR selection were occurring. In two out of three cases, the

frequency of specific OTU’s were significantly different (a = 0.05). While this does not

87

prove that PCR selection is not occurring, it does show that if it was, it was not a
consistent process. In a different study performed at KBS the diversity of denitrifiers
from the CT and NT treatments was assessed using RF LP on PCR amplified nosZ genes
[58]. As in the present study, higher overall diversity, richness and evenness were found
in the communities from the CT treatment. It is unlikely that the same results would be
found using two different genes from the same pathway if PCR bias were occurring to a
large extent.

Having determined that agricultural management influenced denitrifier diversity
and community composition, a possible ecological strategy employed by denitrifiers was
also investigated. There was a significant, positive relationship between dissimilatory
nitrate reducer growth rate and rrn operon copy number. The same relationship was
previously found for aerobic, heterotrophic soil microorganisms, indicating that rm
operon copy number reflects the ecological strategy of microorganisms [23]. Therefore,
there may be a trade-off occurring amongst dissimilatory nitrate reducers between the
capability to respond quickly to inputs of nutrients and the energetic cost of maintaining
multiple ribosomes during stable conditions. The HT and NT treatments both have
significantly higher amounts of total carbon at 0-7 cm (Table 3.3) and available carbon at
0—25 cm than the CT treatment. Due to its higher level of carbon limitation, it could be
argued that the CT treatment is then a more competitive environment, which could select
for organisms with K-selected traits, such as slower, more efficient growth. In that case,
it would be expected that low rrn copy number microorganisms would be favored in the
CT treatment, while high rm copy number organisms would be favored in the less

competitive HT and NT treatments.

88

N20 flux differs significantly between the treatments in that the CT treatment has
an average annual flux that is approximately three times higher than that of the HT and
NT treatments. Besides higher N20 production, other differences have been noted when
agricultural soils are compared to uncultivated soils. Upon cultivation, there is generally
a loss of carbon and subsequent decrease in the C/N ratio [60, 61], along with an increase
in nitrate and decrease in ammonium. These changes result in situations where
denitrifiers experience periods when nitrate is in excess of carbon, and N03' is
incompletely utilized so that N20 mole fraction increases relative to situations when
carbon is not limiting [62, 63]. Supporting this is the fact that previous work at the KBS
LTER indicated activity of 21032 in the NT treatment was significantly higher than in the
CT treatment [8]. NosZ catalyzes the final step in denitrification in which N20 is reduced
to N2, and a decrease in its activity relative to N20 forming enzymes would be indicative
of an increase in N20 mole ratio.

In conclusion, a model is proposed to explain the observed differences in
denitrifier communities and soil physical factors after cultivation (Fig. 3.8). The model
depicts the changes observed in community composition in the three soil treatments at the
KBS LTER which vary in their amount of agricultural intensity. It is proposed that
increases in agricultural intensity result in a change in physical factors so that previously
dominant denitrifiers are lost or decrease in prevalence within the community, causing an
increase in denitrifier diversity, richness and evenness. The dominant denitrifiers present
before agricultural management are those selected for based on their ability to completely
reduce N03‘ to N2. By allowing electron flow through nitrous oxide reductase (Nos)

instead of stopping at nitric oxide reductase (Nor), a second electron sink is present,

89

A.

Agricultural Intensity Gradient

 

 

 

 

 

Low Intensity : A, High Intensity
(HT & NT) (CT)
N20
N20 N20 ? N20
N'lllikc Community Shlft CT-like
Community N20 Community
L - HT-like - Community Shift
-' f Community - .
B. Community Characteristics:
Low Intensity (HT & NT): High Intensity (CT):
1. Dominant HT & NT-like 1. Loss or decrease in NT
OTU's & HT—like OTU's
2. Low richness & evenness 2. High richness & evenness
3. (umpiulc l‘t‘llilt‘liilll “l \U_(' 5. lnrompll'ic l‘t'llllt iinu til \()3’
(\(l5‘ in \3) ll:|\ ni'ul (\( )3‘ in \_-( )) i‘.l\ (H't‘tl
4. High H u (up) 3 lnxorul 4. inn in; (up) A“ i;l\nl‘t‘(i

Figure 3.8: A) Model describing denitrifier community shifts in response to changes in
an agricultural intensity gradient. The dashed line indicates a speculated community shift
from an HT-like community to a NT-like community given enough time. B) The
community characteristics that were noted in this study and also speculated upon. Points
proven in this study are in black; speculated properties of the denitrifier community are in
grey.

90

resulting in faster electron flow and more protons translocated across the cytOplasmic
membrane per unit time. This community is also predicted to be dominated by organisms
with a high rrn operon copy number. After a site is converted to agriculture, the majority
of denitrifiers present will be selected for based on their ability to efficiently utilize
carbon. Efficient carbon use results by stopping denitrification at N20 because at this
point in the reaction, the maximum number of protons have been translocated per
electron. This model is based on the y-proteobacteria Pseudomonas stutzeri which is
capable of translocating a proton when NO is reduced to N20, but not when N20 is
reduced to N2 [1]. This environment should also favor low rrn operon copy number
organisms that will be at an advantage in a competitive, carbon limited environment.
Removal of a site from agriculture allows another community shift to occur in which
denitrifiers capable of complete N03' reduction begin to again dominate the community.
This work, along with that previously conducted at the KBS LTER, demonstrates
that a difference in denitrifier communities along with differences in functional properties
and ecological strategies of the communities may be responsible for the differences in

N20 flux seen in agricultural and successional soils.

91

REFERENCES:

l.

10.

ll.

Zumft, W., Cell biology and molecular basis of denitrrfication. Microbiol. Mol.
Biol. Rev., 1997. 61(4): p. 533-616.

Ferguson, S.J., Nitrogen cycle enzymology. Current Opinion in Chemical Biology,
1998. 2: p. 182-193.

Tiedje, J .M., Ecology of denitrrfication and dissimilatory nitrate reduction to
ammonium, in Biology of Anaerobic Organisms, R. Mitchell, Editor. 1988, John
Wiley & Sons: New York. p. 179-244.

Granli, T. and QC. Bockman, Nitrous oxide from agriculture. Norwegian Journal
of Agricultural Sciences Supplement, 1994. 12: p. 7-128.

Bergsma, T.T., G.P. Robertson, and NE. Ostrom, Influence of soil moisture and
land use history on denitrrfication end-products. Journal of Environmental
Quality, 2002. 31(3): p. 711-717.

Houghton, J.T., Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K.
Maskell, and CA. Johnson, eds. IPCC, 2001: Climate change 2001: the scientific
basis. Contribution of working group I to the third assessment report of the

Intergovernmental panel on climate change. 2001, Cambridge University Press:
Cambridge, United Kingdom and New York, NY, USA. 881.

Matson, P.A., R. Naylor, and I. Ortiz-Monasterio, Integration of environmental,
agronomic, and economic aspects of fertilizer management. Science, 1998.
280(5360): p. 112-115.

Cavigelli, M.A. and G.P. Robertson, The functional significance of denitrifier
community composition in a terrestrial ecosystem. Ecology, 2000. 81(5): p. 1402-
1414.

Cavigelli, M.A. and G.P. Robertson, Role of denitrifier diversity in rates of
nitrous oxide consumption in a terrestrial ecosystem. Soil Biology and
Biochemistry, 2001. 33: p. 297-310.

Holtan-Hartwig, L., P. Dorsch, and LR. Bakken, Low temperature control of soil
denitrifying communities: kinetics of N20 production and reduction. Soil Biology
& Biochemistry, 2002. 34(11): p. 1797-1806.

Rich, J.J., R.S. Heichen, P.J. Bottomley, K. Cromack, Jr., and DD. Myrold,
Community Composition and Functioning of Denitrifying Bacteria from Adjacent
Meadow and Forest Soils. Applied and Environmental Microbiology, 2003.
69(10): p. 5974-5982.

92

12.

13.

14.

15.

16.

l7.

18.

19.

20.

21.

22.

23.

24.

Cheneby, D., S. Perrez, C. Devroe, S. Hallet, Y. Couton, F. Bizouard, G. Iuretig,
J .C. Germon, and L. Philippot, Denitrijying bacteria in bulk and maize-
rhizospheric soil: diversity and N20-reducing abilities. Canadian Journal of
Microbiology, 2004. 50(7): p. 469-474.

Del Grosso, S.J., A.R. Mosier, W.J. Parton, and D.S. Ojima, DAYCENT model
analysis of past and contemporary soil N20 and net greenhouse gas flux for
major crops in the USA. Soil & Tillage Research, 2005. 83(1): p. 9-24.

IPCC, Revised 1 996 Guidelines for National Greenhouse Gas Inventories. 1997.

Huston, M.A., Biological diversity: the coexistence of species on changing
landscapes. 1994, Cambridge: Cambridge University Press. 681.

Buckling, A., R. Kassen, G. Bell, and RB. Rainey, Disturbance and diversity in
experimental microcosms. Nature, 2000. 408(68 l 5): p. 961-964.

Kassen, R., A. Buckling, G. Bell, and RB. Rainey, Diversity peaks at
intermediate productivity in a laboratory microcosm. Nature, 2000. 406(6795): p.
508-512.

Homer-Devine, M.C., M.A. Leibold, V.H. Smith, and B.J.M. Bohannan,
Bacterial diversity patterns along a gradient of primary productivity. Ecology
Letters, 2003. 6(7): p. 613-622.

Connell, J.H., Diversity in Tropical Rain Forests and Coral Reefs - High

Diversity of Trees and Corals Is Maintained Only in a Non-Equilibrium State.
Science, 1978. 199(4335): p. 1302-1310.

Fox, J .F., Intermediate-Disturbance Hypothesis. Science, 1979. 204(4399): p.
1344-1345.

Waide, R.B., M.R. Willig, C.F. Steiner, G. Mittelbach, L. Gough, S.I. Dodson,
G.P. J uday, and R. Parmenter, The relationship between productivity and species
richness. Annual Review of Ecology and Systematics, 1999. 30: p. 257-300.

Groner, E. and A. Novoplansky, Reconsidering diversity-productivity
relationships: directness of productivity estimates matters. Ecology Letters, 2003.
6(8): p. 695-699.

Klappenbach, J .A., J .M. Dunbar, and T.M. Schmidt, rRNA Operon Copy Number
Reflects Ecological Strategies of Bacteria. Appl. Environ. Microbiol., 2000.
66(4): p. 1328-1333.

Coyne, M.S., A. Arunakumari, B.A. Averill, and J.M. Tiedje, Immunological
identification and distribution of dissimilatory heme cd, and nonheme copper

93

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

nitrite reductases in denitrrfi'ing bacteria. Applied and Environmental
Microbiology, 1989. 55(11): p. 2924-2931.

Braker, G., A. Fesefeldt, and K.-P. Witzel, DeveIOpment of PCR primer systems
for amplication of nitrite reductase genes (nirK and nirS) to detect denitrtfling

bacteria in environmental samples. Applied and Environmental Microbiology,
1998. 64(10): p. 3769-3775.

Robertson, G.P., B.A. Paul, and RR. Harwood, Greenhouse gases in intensive

agriculture: contributions of individual gases to the radiative forcing of the
atmosphere. Science, 2000. 289: p. 1922-1925.

Robertson, G.P., B.A. Paul, and RR. Harwood, Greenhouse gases in intensive

agriculture: contributions of individual gases to the radiative forcing of the
atmosphere. Science, 2000. 289: p. 1922-1925.

Gersmehl, P.J., Soil taxonomy and mapping. Annals of the Association of
American Geographers, 1977. 67(3): p. 419-428.

Zhou, J ., M.A. Bruns, and J.M. Tiedje, DNA Recovery from soils of diverse
composition. Applied and Environmental Microbiology, 1996. 62(2): p. 316-322.

Shea, K. and P. Chesson, Community ecology theory as a framework for
biological invasions. Trends in Ecology & Evolution, 2002. 17(4): p. 170-176.

Standard soil methods for long-term ecological research, ed. G.P. Robertson, et
al. 1999, Oxford: Oxford University Press. 462.

Ginard, M., J. Lalucat, B. Tummler, and U. Romling, Genome organization of
Pseudomonas stutzeri and resulting taxonomic and evolutionary considerations.
International Journal of Systematic Bacteriology, 1997. 47(1): p. 132-143.

Hallin, S. and P.-E. Lindgren, PCR detection of genes encoding nitrite reductase
in denitrifying bacteria. Applied and Environmental Microbiology, 1999. 65(4):
p. 1652-1657.

Kim, K.S., T.G. Lilbum, M.J. Renner, and J.A. Breznak, mi] and arflI, Two
genes encoding alpha-L-arabinofuranosidases in Cytophaga xylanolytica.
Applied and Environmental Microbiology, 1998. 64(5): p. 1919-1923.

Ludwig, W., O. Strunk, R. Westram, L. Richter, H. Meier, Yadhukumar, A.
Buchner, T. Lai, S. Steppi, G. Jobb, W. Forster, I. Brettske, S. Gerber, A.W.
Ginhart, 0. Gross, S. Grumann, S. Hermann, R. Jost, A. Konig, T. Liss, R.
Lussmann, M. May, B. Nonhoff, B. Reichel, R. Strehlow, A. Stamatakis, N.
Stuckrnann, A. Vilbig, M. Lenke, T. Ludwig, A. Bode, and K.-H. Schleifer, ARB:

94

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

a software environment for sequence data. Nucl. Acids Res., 2004. 32(4): p.
1363-1371.

Schloss, P.D., B.R. Larget, and J. Handelsman, Integration of Microbial Ecology

and Statistics: a Test T 0 Compare Gene Libraries. Appl. Environ. Microbiol.,
2004. 70(9): p. 5485-5492.

Lozupone, C. and R. Knight, UniFrac: a new phylogenetic method for comparing
microbial communities. Applied and Environmental Microbiology, 2005. 71(12):
p. 8228-8235.

Schloss, PD. and J. Handelsman, Introducing DOTUR, a Computer Program for
Defining Operational Taxonomic Units and Estimating Species Richness. Appl.
Environ. Microbiol., 2005. 71(3): p. 1501-1506.

Colwell, R.K., EstimateS: Statistical estimation of species richness and shared
species from samples. 1994-2004.

Kumar, S., K. Tamura, and M. Nei, MEGA3: An integrated software for
molecular evolutionary genetic analysis and sequence alignment. Integrative and
Comparative Biology, 2003. 43(6): p. 947-947.

Konstantinidis, K.T. and J .M. Tiedje, Genomic insights that advance the species
definition for prokaryotes. Proceedings of the National Academy of Sciences of
the United States of America, 2005. 102(7): p. 2567-2572.

Magurran, A.E., Measuring biological diversity. 2004, Oxford: Blackwell Science
Ltd. 256.

Pagel, M., Inferring evolutionary processes from phylogenies. Zoologica Scripta,
1997. 26(4): p. 331-348.

Pagel, M., Inferring the historical patterns of biological evolution. Nature, 1999.
401(6756): p. 877-884.

Bazylinski, D.A. and RF. Blakemore, Denitrification and Assimilatory Nitrate
Reduction in Aquaspirillum-Magnetotacticum. Applied and Environmental
Microbiology, 1983. 46(5): p. 1118-1124.

Kim, J .K., B.K. Lee, SH. Kim, and J .H. Moon, Characterization of denitrijying
photosynthetic bacteria isolated from photosynthetic sludge. Aquacultural
Engineering, 1999. 19(3): p. 179-193.

Bazylinski, D.A., E. Palome, N.A. Blakemore, and RP. Blakemore,
Denitri/ication by Chromobacterium- Violaceum. Applied and Environmental
Microbiology, 1986. 52(4): p. 696-699.

95

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

Klappenbach, J.A., P.R. Saxman, J.R. Cole, and T.M. Schmidt, rrndb: the
Ribosomal RNA Operon Copy Number Database. Nucleic Acids Research, 2001.
29(1): p. 181-184.

Nehl, DB, 8.]. Allen, and J .F. Brown, Deleterious rhizosphere bacteria: An
integrating perspective. Applied Soil Ecology, 1997. 5(1): p. 1-20.

Siciliano, S.D., C.M. Theoret, J.R. de Freitas, P.J. Hucl, and JJ. Germida,
Diflrerences in the microbial communities associated with the roots of difi’erent

cultivars of canola and wheat. Canadian Journal of Microbiology, 1998. 44(9): p.
844-851.

Grayston, S.J., S.Q. Wang, C.D. Campbell, and A.C. Edwards, Selective influence
of plant species on microbial diversity in the rhizosphere. Soil Biology &
Biochemistry, 1998. 30(3): p. 369-378.

Wieland, G., R. Nerunann, and H. Backhaus, Variation of microbial communities
in soil, rhizosphere, and rhizoplane in response to crop species, soil type, and
crop development. Applied and Environmental Microbiology, 2001. 67(12): p.
5849-5854.

Kent, AD. and E.W. Triplett, Microbial communities and their interactions in
soil and rhizosphere ecosystems. Annual Review of Microbiology, 2002. 56: p.
211-236.

Buckley, DH. and T.M. Schmidt, The structure of microbial communities in soil
and the lasting impact of cultivation. Microbial Ecology, 2001. 42(1): p. 11-21.

Buckley, DH. and T.M. Schmidt, Diversity and dynamics of microbial
communities in soils from agro-ecosystems. Environmental Microbiology, 2003.
5(6): p. 441-452.

Compton, J .E. and RD. Boone, Long-term impacts of agriculture on soil carbon
and nitrogen in New England forests. Ecology, 2000. 81(8): p. 2314-2330.

Knops, J .M.H. and D. Tilrnan, Dynamics of soil nitrogen and carbon
accumulation for 61 years after agricultural abandonment. Ecology, 2000. 81(1):
p. 88-98.

Stres, B., I. Mahne, G. Avgustin, and J.M. Tiedje, Nitrous Oxide Reductase
(nosZ) Gene Fragments Differ between Native and Cultivated Michigan Soils.
Applied and Environmental Microbiology, 2004. 70(1): p. 301-309.

Polz, M.F. and CM. Cavanaugh, Bias in template-to-product ratios in
multitemplate PCR. Applied and Environmental Microbiology, 1998. 64(10): p.
3724-3730.

96

60.

61.

62.

63.

Johnson, D.W., Efl‘ects of Forest Management on Soil Carbon Storage. Water Air
and Soil Pollution, 1992. 64(1-2): p. 83-120.

Davidson, B.A. and LL. Ackerman, Changes in Soil Carbon Inventories
Following Cultivation of Previously Untilled Soils. Biogeochemistry, 1993. 20(3):
p. 161-193.

Firestone, MK. and B.A. Davidson, Microbiological basis of NO and N20
production and consumption in soil, in Exchange of Trace Gases between
Terrestrial Ecosystems and the Atmosphere, M.O. Andreae and D.S. Schimel,
Editors. 1989, John Wiley & Sons: New York. p. 7-21.

Schalk-Otte, S., R.J. Seviour, J .G. Kuenen, and M.S.M. Jetten, Nitrous oxide

(N20) production by Alcaligenes faecalis during feast and famine regimes. Water
Research, 2000. 34(7): p. 2080-2088.

97

CHAPTER 4

THE DIVERSITY OF PLANC T OM YCE TALES IN SOILS DIFFERING IN
HISTORY OF AGRICULTURE

ABSTRACT:

The order Planctomycetales is comprised of diverse, morphologically unique
bacteria that are ubiquitous in nature. Given their potential for important roles in soil
nutrient cycles, especially nitrogen cycling, a phylogenetic survey using 16S rRNA gene
sequences was undertaken to assess soil planctomycete diversity in three soils differing in
their history of agriculture. Sequences clustered with the four genera of planctomycetes
as well as in clusters outside the recognized genera. Three sequences were similar to 16S
rDNA sequences from organisms capable of anaerobic ammonia oxidation, but formed a
separate group. Comparative analyses indicated that two soil treatments that have been
used for agriculture are more similar in community composition and diversity measures
to each other than to a treatment never used for agriculture. However, there were no
significant differences found when comparing the three communities, suggesting that

Planctomycetales are marginally affected by agricultural practices.

INTRODUCTION:

Members of the order Planctomycetales have drawn interest since their discovery
in the early 1900’s [1, 2]. Traits of interest include the recent discovery of the capacity
for anaerobic ammonium oxidation (anammox) by some members of the order and
physiological traits that resemble those of eukaryotes. Despite numerous studies on

aquatic planctomycetes and the recent cultivation of isolates from soil, not much is

98

known about their ecological role in either environment. The information that is known
about planctomycetes however, suggests that they may play important roles in nutrient
cycling.

Historically, it was assumed that planctomycetes grew only in aerobic, aquatic
habitats, as this is where they were commonly cultured. Planctomycetes have been
observed in fresh, marine, and brackish water [3, 4], as well as in aquatic environments
that vary in their trophic status; however they seem to be most prevalent in eutrophic
habitats [4]. Besides aquatic habitats, planctomycetes have been shown to make up a
measurable portion of soil microbial pOpulations (approximately 2 to 11%) by studies
employing rDNA libraries [5-12], rRNA hybridization [13, 14] and FISH [15, 16].
Previously, the order was categorized as aerobic, but there is recent evidence that obligate
anaerobes exist. Isolates have been recovered from an anoxic bioreactor [17], anoxic
sediment [18], an anaerobic wastewater reactor [19], and anoxic rice microcosms [20]. In
addition, Planctomycetales have been isolated from the postlarvae of the Giant Tiger
Prawn, Penaeus monodon [21]. At this time it is not known whether these findings
denote a symbiotic or commensal relationship; however it has been found that a species
of Verrucomicrobia, an order that is one of the closest relations to the Planctomycetales,
is an endosyrnbiont of nematodes of the genus Xiphinema [22]. The fact that
planctomycetes are more ubiquitous, and present in a wider variety of environments than
previously thought suggests that they play an important role in those environments.

Known members of the order Planctomycetales have distinct morphological and
biochemical characteristics that make them unique among the eubacteria and suggest

possible ecological roles. Planctomycetales are the only known cell-wall containing

99

eubacteria besides the Chlamydiae and mycoplasmas that lack peptidoglycan [23-25]. In
a study to identify a medium suitable for the isolation of aquatic Planctomycetales, N-
acetylglucosamine was used as both a carbon and nitrogen source [4]. N-
acetylglucosamine is a component of peptidoglycan, and its consumption by an organism
lacking this same compound implies that the Planctomycetales have a role in the
degradation or mineralization of cell wall material. Also, some planctomycetes produce
stalks that can act as hold-fasts, allowing the organisms to attach to substrates in the
environment or each other [3]. Manganese and iron oxide encrustations have been found
on the stalks of some aquatic Planctomycetales leading to the speculation that they may
be capable of manganese and iron oxidation [26, 27].

One of the most intriguing characteristics of the Planctomycetales is the recent
discovery that some members play an important role in global nitrogen cycling.
Uncultured representatives have been shown to carry out anaerobic oxidation of ammonia
(anammox) [28]. During anammox, ammonia is oxidized and nitrate or nitrite reduced to
form dinitrogen gas [29, 30]. This process has been detected in wastewater treatment
plants, freshwater and marine sediments, and the anoxic ocean water column. Recently it
was estimated that anammox is responsible for up to 50% of the removal of fixed
nitrogen from the ocean [31, 32]. There have been no reports of anammox being detected
in soil, but a full investigation into its occurrence has not been reported.

Recent work at the Kellogg Biological Station Long-Term Ecological Research
(KBS LTER) site has shown that soils differing in inorganic and total nitrogen as well as
carbon and history of agricultural use, also differ in their amounts of nitrous oxide flux

[33]. This indicates that different microbial communities involved in nitrogen cycling

100

may be present at each site. Due to the possible involvement of the Planctomycetales in
nitrogen cycling, an investigation was performed in which the diversity and community
composition of planctomycetes from three soil treatments were compared. A
phylogenetic analysis was also performed to determine how closely soil organisms were

related to anammox organisms.

MATERIALS & METHODS:
Study Site & Soil Collection

Soil samples were collected in October 1996 from the Kellogg Biological Station
Long-Term Ecological Research Site (KBS LTER) located in Hickory Comers, MI. The
dominant soil series at the station are the Kalamazoo and Oshtemo series. These are fine-
loarny and coarse-loamy mesic Typic Hapludalfs, respectively. The KBS LTER has soil
treatments that endure a wide range of human impact. Three treatments were focused on
in this study; a conventional agriculture (CT) site that receives amounts of fertilizer,
herbicide, and tillage that are typical for the Midwest region and is on an annual corn
(Zea mays L.) - soybean (Glycine max L.) - wheat (T riticum aestivum L.) rotation, a
historically tilled site (HT) abandoned from agriculture in 1989, and a never-tilled site
(NT). The KBS LTER is described fully on the World Wide Web at http://
lter.kbs.msu.eduk

Soil was collected at a depth of 0-10 cm fiom the CT, HT and NT treatments. At
each treatment plot, a total of five cores of a 2.5 cm diameter were taken, and cores from

the same treatment were pooled, sieved (2 mm mesh), and frozen in liquid nitrogen in the

101

field. The soil was kept on dry ice until it was brought back to the laboratory and stored
at -80°C.
Library Construction

One library was made from DNA extracted from each of the three soil treatments.
Total soil DNA was extracted from 0.25 - 1.0 g soil using an UltraClean Soil DNA Kit
(MoBio, Carlsbad, CA). The Planctomycetales specific forward primer Pla37F (5’ TGG
CGG CRT GGA TTA G 3’; modified fi'om Neef and colleagues [34]) and universal
reverse primer 1540R (5’ AAG GAG GTG ATC CAR CCG CA 3’) were used to amplify
16S rRNA genes (both primers modified or designed by Daniel Buckley, personal
communication). PCR amplification was performed with 1 U of Amplitaq Gold® Taq
polymerase (Applied Biosystems, Foster City, CA), a 1X concentration of the
manufacturer’s PCR buffer (10 mM Tris-HCl [pH 8.3], 50 mM KCl, 1.5 mM MgCl2,
0.001% (w/v) gelatin), 2.0 mM MgCl2, 0.2 mM dNTP’s (Roche, Indianapolis, IN), 0.01%
BSA, 50 pmol of each primer, and 10-50 ng of template DNA in a 25 [LL volume. PCR
mixtures were incubated in a GeneAmp PCR System 9600 therrnocycler (Perkin-Elmer,
Boston, MA) with the following PCR protocol: (i) 12 min initial denaturation at 950°C;
(ii) 30 cycles of denaturation for 30 s at 950°C, primer annealing for 403 at 64.0°C,
extension for 45 s at 720°C; and (iii) a final extension for 10 min at 720°C. Genomic
DNA from Planctomyces limnophilus ATCC 43296 was used as a positive control and
Verrucomicrobiurn spinosum ATCC 43997 as a negative control for the PCR.

Each 168 rDNA clone library was created using three pooled reactions. Purified
product was cloned using a TOPO TA Cloning Kit for Sequencing with cloning vector

pCR 2.1 and transformed into Escherichia coli One Shot TOP10® competent cells

102

(Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Transformants
were plated on LB agar containing 100 ug/mL ampicillin. Clones were screened for the
correct insert size and the PCR product sequenced using a model 373A DNA sequencer
(Applied Biosystems) with dye-terminator fluorescent cycle sequencing technology. All
sequences were checked with the Chimera Check tool [35] to ensure that no chimeric
sequences were included in the analysis. A total of 98 sequences were found to be of
sufficient quality to be included in phylogenetic analyses, during which three were found
to group more closely to sequences from Chlamydia and Verrucomicrobia than
planctomycetes. Therefore, the three sequences were not included in further community
composition or diversity analyses. This resulted in there being 43 clones in the CT
library, 27 in the HT library, and 25 in the NT library. Full-length sequencing of seven
16S rRNA genes was completed to aid in phylogenetic analyses.
Phylogenetic Analyses

Phylogenetic analyses were performed using the program ARB [36]. Previously
published Planctomycetales l6S rDNA sequences were obtained from NCBI and aligned
along with the sequences obtained in this study. Sequence alignments were performed
using the automatic aligner in ARB followed by visual inspection and correction as
needed. Sections of the sequence with ambiguous alignments as well as the primer
sequences were excluded from the analyses. The phylogenetic tree was constructed using
the neighbor joining method.
Diversity Analyses

A nucleotide similarity cutoff of 97 % was used to define Planctomycetales

operational taxonomic units (OTU’s). This cutoff was chosen due to the fact that it was

103

shown to correspond well with the traditional 70% DNA-DNA reassociation standard
currently used to define microbial species [37].

l-LIBSHUFF version 1.3 [38] was used to analyze res rDNA libraries to
determine whether those created fi'om different treatments were statistically different,
while DOTUR version 1.53 [39] was used to analyze the a-diversity of each library.
Input into DOTUR consisted of neighbor joining distance matrices from each individual
library. DOTUR output was used to calculate Simpson’s diversity indices (-lnD),
Simpson’s evenness measure ([1/D]/S, where S is equal to the number of OTU’s present
in each library rarified to the number of clones in the smallest library), Good’s coverage
([1-(n/N)]*100, where n equals the number of singletons and N equals the total number
of sequences), and Chaol curves at a 97% nucleotide similarity cutoff.

In addition, B-diversity was determined using both DOTUR and l-LIBSHUFF.
DOTUR was used to separate sequences from all libraries into OTU’s and the DOTUR
.list file converted into a matrix used as input for the program Estimates version 7.00
[40]. A J accard dissimilarity matrix created with Estimates data was converted to a

dendrograrn using the program MEGA version 3.1 [41].

RESULTS:
Phylogenetic Analysis

Phylogenetic analysis revealed that the 95 cloned l6S rRNA genes fell into four
recognized genera within the order Planctomycetales and also into four groups not
associated with the known genera (Fig. 4.1). Ten sequences clustered with the genus

Pirellula, 11 with the genus Planctomyces, 33 with the genus Gemmata, and 10 with the

104

Group 1 (3)
1 /2/0

  
 

  

HT

 

 

    
 

clone CT Clone Clone
clone A-2 Sva0503 Gm2g3100)
Candidatus HT
anammox clone
organisms
Clone
Sva0500 0‘
Verrucomicrobia _.,,'” ,2;- .
f _ . " Pirellula (10)
Chlamydia ... 3/4/3
Group 3 (l l)
.1 . ‘40] 8,2] I
3’”
pr
ii“ Group 4 (7)
0/5/2

Isosphaera (10)
5/0/5

Planctomyces (l l)

Gemmata (33) 6/4/1

l4/7/12

0.10

Figure 4.1: Neighbor joining phylogenetic tree constructed fi'om full and partial 16S
rDNA sequences. Numbers in parentheses refer to the number of KBS LTER clones
within each group. Numbers separated by backslashes indicate the number of clones in
each group from a specific soil treatment library (conventional agriculture/abandoned
from agriculture/never tilled). Clones in bold are those clustering most closely with
planctomycetes capable of anammox. The scale bar represents a 10% difference between
nucleotide sequences.

105

genus Isosphaera. The majority of the remaining sequences clustered into four groups
which for the purposes of this chapter were designated Groups 1 through 4. Three
sequences clustered in Group 1, 10 clustered in Group 2, 11 clustered in Group 3, and 7
sequences clustered in Group 4. Sequences from each of the three treatments were
present in most of the genera and groups. The exceptions were that no Group 1
sequences from the never tilled treatment, no Isosphaera sequences from the treatment
abandoned from agriculture, and no Group 4 sequences from the conventional agriculture
treatment were found (Fig. 4.1). The 3 clones in Group 1 clustered near 16S rDNA
sequences from Planctomycetales known to perform anammox and showed between 78.8
and 84.6% sequence similarity to the anammox planctomycete sequences.
Community Composition

The composition of the three Planctomycetales soil communities were compared
by statistical comparison of the 16S rRNA gene libraries. Analysis with the program
I-LIBSHUFF showed that overall, none of the communities were significantly different
from each other (Fig. 4.2). Library comparisons in l-LIBSHUFF are performed so that
not only is Replicate Library X compared to Replicate Library Y, but Replicate Y is also
compared to Replicate X. This can result in cases where in one comparison libraries are
significantly different, but in the other they are not. This indicates that one library is a
subset of the other and subsequently the libraries were considered to not be significantly
different. The library from the conventional agriculture treatment was not significantly
different from those from the treatment abandoned fi'om agriculture or the treatment that
had never been tilled. The never tilled treatment library was a subset of the treatment

abandoned fi'om agriculture. In order to ascertain which libraries were more similar to

106

 

CT

 

 

 

 

0.0033 *

HT

0.9321

 

Figure 4.2: Results of l—LIBSHUFF analysis for Planctomycetales 16S rDNA sequence
libraries from three KBS LTER soil treatments (libraries are represented by blocks).
Comparisons between two individual libraries with l—LIBSHUFF actually entail two
statistical comparisons: library X is compared to library Y and Y to X. As such, arrows
point to the library that is being compared to the library being pointed from. Numbers
are p-values from comparisons between libraries, and an asterisk (*) designates p-values
that are significant at a = 0.05.

107

each other, a dendrogram was constructed using a J accard dissimilarity matrix (Fig. 4.3).
Even though none of the communities differ significantly, those from both treatments that
have been used for agriculture (CT and HT) are more similar to each other than the
community from the treatment never used for agriculture (NT).

Diversity of Soil Planctomycetales

Overall diversity, richness, and evenness were calculated for the three soil
treatment libraries (Table 4.1). Simpson’s diversity index was highest in the two
treatments that have been used for agriculture (CT and HT) and lowest in the never tilled
treatment. In order to determine if one of the two components of diversity had a stronger
influence than the other, richness and evenness were compared.

The richness of libraries that differ in number of clones they contain can be
compared by using rarefaction to estimate how many species would be present if all
libraries had as many clones as the smallest library [42]. Rarefaction indicated that the
number of species in each treatment were approximately equal (Table 4.1). Richness can
also be compared by using nonparametric statistics, such as the Chaol richness estimator
[43]. Chaol curves were plotted along with 95% confidence intervals to both compare
richness and assess whether more sampling would be required to obtain an accurate
richness estimate (Fig. 4.4). As with the rarefaction analysis, the Chaol curves indicated
that richness of Planctomycetales in the three soil treatments did not differ. One
interesting point, however, is that the Chaol curves for the conventional agriculture
treatment and the treatment abandoned from agriculture are asymptotic, indicating that no
further sampling is required for an accurate prediction of richness. The curve fi'om the

never tilled treatment library is non-asymptotic, indicating that the clone library is too

108

 

 

 

 

 

 

CT
HT
NT

.L

0.1

Figure 4.3: Dendrogram based on a Jaccard dissimilarity matrix calculated at a 97%
nucleotide similarity cutoff. Scale represents percent dissimilarity.

109

0000088808 .00 800808888 3800 u 2 000 088080350 .00 8000-88888 u 0 080883 608 LEE-a 00 0080—86000 o
@0503 00 0080—88200 8.

0088080 mm 00 8008.08.88 0.3.8.0 .«0 80088088888 0 m 00 8008088808800 o

.QEI 00 0208880800 0

088088.090 00:88 .896: u .82 0080 683888050 80b 0088008800 888008000 0 E 4008.58.88 088888050 80080088080000 0 H0 .

 

 

00 8.0 00 00.8. 808.0 .008 088 00 .02
08 0.0 00 08.0 8000 .000 008 00 .E
mm m0 mm 80.0 88 .NE m8: 9 H0
0.80 .0000
A85 03.808000 00088880>m 505 380.8205 a 008805
e 8V 0. D 80 .80 8008882 0 . . 080803 0880850089
00.000 0.000908 o 0.0008858 0 8 E: a

 

b08080 bra—0050 00880088000 08.0.0 0 880 008.0080: <29 m3 80.8.8 0080888008 00880000802800.0203 80.0 00008880 b80835 88.84 030.8-

110

N U
M O
o o
r r

Chaol Richness Estimator
N
8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Number of Clones

I-;F1pibflp ‘
T-i-i': Fifi"
.«l- ":5
ii is
.3
i 4" i i:
*l l ,
*“l i :l
T5} ‘ Hi?
i l 9.
T: 3 l
I l
I.
1.. 9 -".0J.
7 ”.'O'.'. J
? '0’.’... £- +‘dbdbdbabdhunhbbdb -
10 20 30 40 50

Figure 4.4: Chaol estimates of Planctomycetales richness in three KBS LTER soil
treatments at a 97% 16S rDNA sequence similarity cutoff. (I) Conventional agricultural
treatment, (A) treatment abandoned from agriculture, and (0) never tilled treatment.
Error bars represent 95% confidence intervals.

111

small to obtain an accurate prediction of Planctomycetales richness in this treatment. At
best, the value obtained is a minimum estimate of the richness in this treatment,
suggesting that richness in the never tilled treatment may actually be higher than in the
other two treatments.

Evenness describes the relative abundance of OTU’s in each library and was
highest in the conventional agriculture treatment and the treatment abandoned from
agriculture. This was further illustrated by plotting rank/abundance curves for each
library (Fig. 4.5) and comparing them statistically. Kolomogorov-Smimov 2-sample
tests indicated that the rank/abundance curves of the conventional agriculture and never
tilled treatment differed significantly (P < 0.05). Therefore, differences in evenness have

the largest influence on the Planctomycetales diversity measures.

DISCUSSION:

The order Planctomycetales encompasses a group of organisms with a high level
of diversity as well as unique biochemical and morphological characteristics. Members
of the order are ubiquitous in soils, however not much is known regarding potential
ecological roles of these organisms. In fact, only recently has a group within the order
been identified as being capable of anaerobic oxidation of ammonia, which plays an
important role in ocean nitrogen cycling [31, 32]. Due to the potential of
Planctomycetales to play important, but as yet undiscovered roles in soil nutrient cycles,
a phylogenetic analysis and comparison of community structure in three different soil

treatments was undertaken. The goal of the study was to gain insights into factors driving

112

0.12~
0.10-
0.08-

0.06 -

0.04 4 999m...

Relative Abundance of OTU

0.02 _ _—II-I-I-__I-

 

0.00

 

I r I I ' I ' I ' I

0.5.1015'2'0'25 30 35 40
Rank

Figure 4.5: Rank/abundance curves of Planctomycetales OTU’s defined at a 97%
sequence similarity cutoff. (I) Conventional agricultural treatment, (A) treatment
abandoned from agriculture, and (O) never tilled treatment. Error bars represent 95%
confidence intervals.

113

planctomycetes community structure and to determine whether members of the soil
community have any relation to planctomycetes capable of anammox.

Of the 98 16S rRNA gene clones analyzed, three clustered near the 16S rRNA
gene sequences of organisms capable of anammox. The three soil sequences had percent
nucleotide similarities to the anammox organism sequences of ca. 79 to 85%, which is
intriguing considering that the different genera of anammox capable planctomycetes
show <85% 16S rDNA nucleotide sequence similarity to each other. [44] The clone
libraries constructed from each soil treatment were small and had coverage values of only
15 to 28% (Table 4.1), indicating that further sequencing may result in discovery of
sequences with higher similarity to anammox organism sequences. Anammox
planctomycete-specific primers have been developed and used in aquatic environments
[45] which if applied to soil could reveal the presence of these organisms. In addition,
direct detection of anammox in soil may be possible with the use of stable isotopes as in
recent aquatic studies [31, 45-49].

The three soil treatments focused on were chosen due to their different
agricultural histories. Both a treatment subjected to conventional agricultural practices
and one never used for agriculture were sampled, as well as a treatment that had not been
subjected to agriculture in 7 years at the time soil samples were taken. a- as well as B-
diversity measures revealed that the two treatments with a history of agriculture were
more similar to each other with respect to commmrity composition, diversity, evenness,
and estimated richness than to the treatment never used for agriculture. The fact that the
two treatments with an agricultural history have more similar communities is not

surprising given that previous work has shown that the impact of agricultural practices on

114

 

1rd

microbial communities can last for decades [13, 14]. This is thought to be due to the long
amount of time needed for soil carbon and nitrogen to recover to pre-agricultural
concentration and quality [50, 51].

Despite the fact that the communities of two treatments with a history of
agriculture were more similar to each other than to a treatment never used for agriculture,
there was no significant difference in libraries fi'om the three soils. Initially, this was
unexpected considering that Buckley and Schmidt [13, 14] previously reported that
microbial community composition of the KBS LTER conventional agriculture and
abandoned fi'om agriculture treatments were not significantly different, but did differ
significantly from the never tilled treatment community. However, when the
composition of the Planctomycetales communities were specifically focused on [14], soil
treatment did not have a significant effect on community composition which is in
agreement with the results presented here. This suggests that Planctomycetales are not
contributing to differences seen in total microbial community composition, may only be
marginally affected by agricultural practices, and/or be capable of faster recovery after

cultivation than other microbial groups.

CONCLUSION:

The findings presented in this chapter invite further studies into factors affecting
Planctomycetales community structure and diversity. In particular, the presence of 16S
rDNA sequences clustering near those of anammox organisms, despite the small number
of clones analyzed, indicates that further investigation into the presence of anammox

organisms and the occurrence of anammox in soil is warranted. In addition, preliminary

115

results show that treatment effects are not reflected in planctomycete community
composition, suggesting that the community is only marginally affected by agricultural

practices.

116

REFERENCES:

l.

10.

Girnesi, N., Hydrobiologiai T anulmanyok (Hydrobiologische Studien). I:
Planctomyces bekefii Gim. nov. gen. et sp. [in Hungarian, with German
translation]. Kiadj a a Magyar Ciszterci Rend Budapest, Hungary, 1924: p. 1-8.

Henrici, AT. and DE. Johnson, Studies of Freshwater Bacteria: II. Stalked
Bacteria, a New Order of Schizomycetes J. Bacteriol., 1935. 30(1): p. 61-93.

Staley, J.T., J .A. Fuerst, S. Giovannoni, and H. Schlesner, The order
Planctomycetales and the genera Planctomyces, Pirellula, Gemmata, and

Isosphaera, in The Prokaryotes, A. Ballows, et 31., Editors. 1992, Springer-
Verlag: New York. p. 3710-3731.

Schlesner, H., The development of media suitable for the microorganisms
morphologically resembling Planctomyces spp., Pirellula spp., and other
Planctomycetales from various aquatic habitats using dilute media. Systematic
and Applied Microbiology, 1994. 17: p. 135-145.

Bomeman, J., P.W. Skroch, K.M. Osullivan, J.A. Palus, N.G. Rumjanek, J.L.
Jansen, J. Nienhuis, and E.W. Triplett, Molecular microbial diversity of an

agricultural soil in Wisconsin. Applied and Environmental Microbiology, 1996.
62(6): p. 1935-1943.

Bomeman, J. and E.W. Triplett, Molecular microbial diversity in soils from
eastern Amazonia: Evidence for unusual microorganisms and microbial
population shifts associated with deforestation. Applied and Environmental
Microbiology, 1997. 63(7): p. 2647-2653.

Kuske, C.R., S.M. Barns, and J .D. Busch, Diverse uncultivated bacterial groups
from soils of the arid southwestern United States that are present in many
geographic regions. Applied and Environmental Microbiology, 1997. 63(9): p.
3614-3621.

Lee, S.-Y., J. Bollinger, D. Bezdicek, and A. Ogram, Estimation of the abundance
of an uncultured soil bacterial strain by a competitive quantitative PCR method.
Applied and Environmental Microbiology, 1996. 62(10): p. 3787-3793.

Liesack, W. and E. Stackebrandt, Occurrence of Novel Groups of the Domain
Bacteria as Revealed by Analysis of Genetic Material Isolated from an Australian
Terrestrial Environment. Journal of Bacteriology, 1992. 174(15): p. 5072-5078.

McCaig, A.E., L.A. Glover, and J .I. Prosser, Molecular analysis of bacterial

community structure and diversity in unimproved and improved upland grass
pastures. Applied and Environmental Microbiology, 1999. 65(4): p. 1721-1730.

117

11.

12.

13.

14.

15.

l6.

17.

18.

19.

20.

21.

Ueda, T., Y. Suga, and T. Matsuguchi, Molecular Phylogenetic Analysis of a Soil
Microbial Community in a Soybean Field. European Journal of Soil Science,
1995. 46(3): p. 415-421.

Zhou, J.Z., M.E. Davey, J.B. Figueras, E. Rivkina, D. Gilichinsky, and J.M.
Tiedje, Phylogenetic diversity of a bacterial community determined from Siberian
tundra soil DNA. Microbiology-Uk, 1997. 143: p. 3913-3919.

Buckley, DH. and T.M. Schmidt, The structure of microbial communities in soil
and the lasting impact of cultivation. Microbial Ecology, 2001. 42(1): p. 11-21.

Buckley, DH. and T.M. Schmidt, Diversity and dynamics of microbial
communities in soils from agro-ecosystems. Environmental Microbiology, 2003.
5(6): p. 441-452.

Zarda, B., D. Hahn, A. Chatzinotas, W. Schonhuber, A. Neef, R.I. Amann, and J.
Zeyer, Analysis of Bacterial community Sturcture in Bulk Soil by in situ
Hybridization. 168, 1997: p. 185-192.

Chatzinotas, A., R.A. Sandaa, W. Schonhuber, R. Amann, F .L. Daae, V. Torsvik,
J. Zeyer, and D. Hahn, Analysis of broad-scale drfl'erences in microbial

community composition of two pristine forest soils. Systematic and Applied
Microbiology, 1998. 21(4): p. 579-587.

Godon, J .J ., E. Zumstein, P. Dabert, F. Habouzit, and R. Moletta, Molecular
microbial diversity of an anaerobic digestor as determined by small-subunit

rDNA sequence analysis. Applied and Environmental Microbiology, 1997. 63(7):
p. 2802-2813.

Miskin, I.P., P. Farrimond, and I.M. Head, Identification of novel bacterial
lineages as active members of microbial populations in a fieshwater sediment
using a rapid RNA extraction procedure and R T -PCR. Microbiology-Uk, 1999.
145: p. 1977-1987.

Strous, M., J.G. Kuenen, and SM. Jetten, Key physiology of anaerobic
ammonium oxidation. Applied and Environmental Microbiology, 1999. 65(7): p.
3248-3250.

Derakshani, M., T. Lukow, and W. Liesack, Novel bacterial lineages at the
(sub)division level as detected by signature nucleotide-targeted recovery of 16S
rRNA genes from bulk soil and rice roots of flooded rice microcosms. Applied and
Environmental Microbiology, 2001. 67(2): p. 623-631.

Fuerst, J.A., H.G. Gwilliam, M. Lindsay, A. Lichanska, C. Belcher, J .E. Vickers,
and P. Hugenholtz, Isolation and molecular identification of planctomycete

118

22.

23.

24.

25.

26.

27.

28.

29.

30.

bacteria fiom postlarvae of the Giant Tiger Prawn, Penaeus monodon. Applied
and Environmental Microbiology, 1997. 63(1): p. 254-262.

Vandekerckhove, T .T.M., A. Willems, M. Gillis, and A. Coomans, Occurence of
novel verrucomicrobial species, endosymbiotic and associated with
parthenogenesis in Xiphinema americanum-group species (Nematoda,
Longidoridae). International Journal of Systematic and Evolutionary
Microbiology, 2000. 50: p. 2197-2205.

Kfinig, E., H. Schlesner, and P. Hirsch, Cell wall studies on budding bacteria of
the Planctomyces/Pasteuria group and on a Prosthecomicrobium sp. Archives of
Microbiology, 1984. 138: p. 200-205.

Liesack, W., H. Kbnig, H. Schlesner, and P. Hirsch, Chemical composition of the
peptidoglycan-free cell envelopes of budding bacteria of the Pirella/Planctomyces
group. Archives of Microbiology, 1986. 145: p. 361-366.

Stackebrandt, E., U. Wehmeyer, and W. Leisack, I6S ribosomal RNA- and cell
wall analysis of Gemmata obscuriglobus, a new member of the order
Planctomycetales. FEMS Microbiology Letters, 1986. 37: p. 289-292.

Schmidt, J .M., W.P. Sharp, and MP. Starr, Manganese and Iron encrustiations
and other features of Planctomyces crassus Hortobdgyi 1965, Morphotype lb of
the Blastocaulis-Planctomyces group of budding and appendaged bacteria,
examined by electron microscopy and X-ray micro-analysis. Current
Microbiology, 1981. 5: p. 241-246.

Schmidt, J.M., W.P. Sharp, and MP. Starr, Metallic-oxide encrustations of the
nonprosthecate stalks of naturally occurring papulations of Planctomyces bekefii.
Current Microbiology, 1982. 7: p. 389-394.

Strous, M., J .A. Fuerst, E.H.M. Kramer, S. Logemann, G. Muyzer, K.T. van de
Pas-Schoonen, R. Webb, J .G. Kuenen, and SM. J etten, Missing Lithotroph
Identified as New Planctomycete. Nature, 1999. 400: p. 446-449.

Guven, D., A. Dapena, B. Kartal, M.C. Schmid, B. Maas, K. van de Pas-
Schoonen, S. Sozen, R. Mendez, H.J.M. Op den Camp, M.S.M. Jetten, M. Strous,
and I. Schmidt, Propionate oxidation by and methanol inhibition of anaerobic
ammonium-oxidizing bacteria. Applied and Environmental Microbiology, 2005.
71(2): p. 1066-1071.

Strous, M., E. Pelletier, S. Mangenot, T. Rattei, A. Lehner, M.W. Taylor, M.
Horn, H. Daims, D. Bartol-Mavel, P. Wincker, V. Barbe, N. Fonkrrechten, D.
Vallenet, B. Segurens, C. Schenowitz-Truong, C. Medigue, A. Collingro, B. Snel,
B.E. Dutilh, H.J.M. Op den Camp, C. van der Drift, I. Cirpus, K.T. van de Pas-
Schoonen, H.R. Harhangi, L. van Niftrik, M. Schmid, J. Keltjens, J. van de

119

31.

32.

33.

34.

35.

36.

37.

38.

Vossenberg, B. Kartal, H. Meier, D. Frishman, M.A. Huynen, H.W. Mewes, J.
Weissenbach, M.S.M. Jetten, M. Wagner, and D. Le Paslier, Deciphering the

evolution and metabolism of an anammox bacterium from a community genome.
Nature, 2006. 440(7085): p. 790-794.

Kuypers, M.M.M., G. Lavik, D. Woebken, M. Schmid, B.M. Fuchs, R. Amann,
B.B. J orgensen, and M.S.M. J etten, Massive nitrogen loss from the Benguela
upwelling system through anaerobic ammonium oxidation. Proceedings of the
National Academy of Sciences of the United States of America, 2005. 102(18): p.
6478-6483.

Arrigo, K.R., Marine microorganism and global nutrient cycles. Nature, 2005.
437(7057): p. 349-355.

Robertson, G.P., B.A. Paul, and RR. Harwood, Greenhouse gases in intensive
agriculture: contributions of individual gases to the radiative forcing of the

atmosphere. Science, 2000. 289: p. 1922-1925.

Neef, A., R. Amann, H. Schlesner, and K.-H. Schleifer, Monitoring a widespread
bacterial group: in situ detection of planctomycetes with I 6S rRNA -targeted
probes. Microbiology, 1998. 144: p. 3257-3266.

Cole, J.R., B. Chai, R.J. Farris, Q. Wang, S.A. Kularn, D.M. McGarrell, G.M.
Garrity, and J .M. Tiedje, The Ribosomal Database Project (RDP-II): sequences
and tools for high-throughput rRNA analysis. Nucl. Acids Res., 2005.
33(suppl_l): p. D294-296.

Ludwig, W., O. Strunk, R. Westram, L. Richter, H. Meier, Yadhukumar, A.
Buchner, T. Lai, S. Steppi, G. Jobb, W. Forster, I. Brettske, S. Gerber, A.W.
Ginhart, 0. Gross, S. Grumann, S. Hermann, R. Jost, A. Konig, T. Liss, R.
Lussmann, M. May, B. Nonhoff, B. Reichel, R. Strehlow, A. Stamatakis, N.
Stuckrnann, A. Vilbig, M. Lenke, T. Ludwig, A. Bode, and K.-H. Schleifer, ARB:
a software environment for sequence data. Nucl. Acids Res., 2004. 32(4): p.
1363-1371.

Stackebrandt, E. and BM. Goebel, A Place for DNA-DNA Reassociation and 16S
Ribosomal-RNA Sequence-Analysis in the Present Species Definition in
Bacteriology. International Journal of Systematic Bacteriology, 1994. 44(4): p.
846-849.

Schloss, P.D., B.R. Larget, and J. Handelsman, Integration of Microbial Ecology

and Statistics: 0 Test T 0 Compare Gene Libraries. Appl. Environ. Microbiol.,
2004. 70(9): p. 5485-5492.

120

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

Schloss, PD. and J. Handelsman, Introducing DOTUR, a Computer Program for
Defining Operational Taxonomic Units and Estimating Species Richness. Appl.
Environ. Microbiol., 2005. 71(3): p. 1501-1506.

Colwell, R.K., EstimateS: Statistical estimation of species richness and shared
species from samples. 1994-2004.

Kumar, S., K. Tamura, and M. Nei, MEGA3: An integrated software for
molecular evolutionary genetic analysis and sequence alignment. Integrative and
Comparative Biology, 2003. 43(6): p. 947-947.

Magurran, A.E., Measuring biological diversity. 2004, Oxford: Blackwell Science
Ltd. 256.

Chao, A., Nonparametric-Estimation of the Number of Classes in a Population.
Scandinavian Journal of Statistics, 1984. 11(4): p. 265-270.

Jetten, M.S.M., I. Cirpus, B. Kartal, L. van Niftrik, K.T. van de Pas-Schoonen, O.
Sliekers, S. Haaijer, W. van der Star, M. Schmid, J. van de Vossenberg, I.
Schmidt, H. Harhangi, M. van Loosdrecht, J.G. Kuenen, H.O. den Camp, and M.
Strous, 1994-2004: 10 years of research on the anaerobic oxidation of
ammonium. Biochemical Society Transactions, 2005. 33: p. 119-123.

Tal, Y., J .E.M. Watts, and H.J. Schreier, Anaerobic ammonia-oxidizing bacteria
and related activity in Baltimore inner Harbor sediment. Applied and
Environmental Microbiology, 2005. 71(4): p. 1816-1821.

Thamdrup, B. and T. Dalsgaard, Production of N-2 through anaerobic ammonium
oxidation coupled to nitrate reduction in marine sediments. Applied and
Environmental Microbiology, 2002. 68(3): p. 1312-1318.

Dalsgaard, T., D.E. Canfield, J. Petersen, B. Thamdrup, and J. Acuna-Gonzalez,
N-2 production by the anammox reaction in the anoxic water column of Golfo
Dulce, Costa Rica. Nature, 2003. 422(6932): p. 606-608.

Risgaard-Petersen, N., R.L. Meyer, M. Schmid, M.S.M. Jetten, A. Enrich-Prast,
S. Rysgaard, and NP. Revsbech, Anaerobic ammonium oxidation in an estuarine
sediment. Aquatic Microbial Ecology, 2004. 36(3): p. 293-304.

Rysgaard, S., RN. Glud, N. Risgaard-Petersen, and T. Dalsgaard, Denitrtfication
and anammox activity in Arctic marine sediments. Limnology and Oceanography,
2004. 49(5): p. 1493-1502.

Compton, J .E. and RD. Boone, Long-term impacts of agriculture on soil carbon
and nitrogen in New England forests. Ecology, 2000. 81(8): p. 2314-2330.

121

51. Knops, J.M.H. and D. Tilrnarr, Dynamics of soil nitrogen and carbon
accumulation for 61 years after agricultural abandonment. Ecology, 2000. 81(1):
p. 88-98.

122

CHAPTER 5

CONCLUSIONS AND FUTURE DIRECTIONS

SIGNIFICANCE:

Microbial communities are essential players in agricultural soil nutrient cycles as
their activity helps regulate the availability of nutrients to crops. Despite this important
role, the main factors affecting microbial diversity and community structure are just
beginning to be uncovered. As well as identifying factors affecting microbial
communities, it is important to determine the links between particular communities and
the ecosystem functions that they perform. The identification and understanding of
microbial communities and their interactions will allow more accurate predictions to be
made regarding the effect of agricultural practices on ecosystem functions. For example,
the work presented in this dissertation and fi'om recent published studies indicates that
greenhouse gas flux cannot be explained solely by physical factors and suggests that
differences in microbial populations should be accounted for in flux models in order to
improve their accuracy.

The focus of the work presented in this dissertation was on microbial populations
known to be involved in soil nitrogen cycling (denitrifiers), or with the potential for
involvement (Planctomycetales). Denitrification in particular is detrimental in
agricultural soils as it leads to loss of added nitrogen and production of the greenhouse
gas, N2O. Therefore, statistical modeling (Chapter 2) and molecular methods (Chapters 3

and 4) were used to investigate how physical aspects of the environment affect microbial

123

communities and, in turn, how microbial communities may affect the ecosystem function
OszO flux.
A summary of the work presented in this dissertation is listed below and followed

by suggestions for further research.

SUMMARY:

Chapter 2:

1. Field measurements of physical factors affecting microbial populations explained
between 8 and 50% of the variation in C02, CH4 and N20 flux at the KBS LTER,
with CO2 flux consistently having the greatest amount of explainable variation
(29 to 50%). Since CO2 is a byproduct of all types of heterotrophic respiration in
soil, its flux should be responsive to factors enhancing microbial metabolism in
general, and this explained the high coefficients of determination obtained for
CO2 flux. Conversely, since the consumption of CH4 and production of N20
require the activity of microbes with specialized metabolic pathways;
environmental parameters alone explain a smaller percentage of the flux of these
gases.

2. In plots used for conventional agriculture, there were significant effects on gas
flux associated with specific crops, with fluxes roughly following the pattern
wheat > soybean > com. This indicates that the annual rotation of crops at KBS
may result in annual changes in microbial communities that vary in their

metabolic capabilities and subsequent gas flux.

124

The treatment with the highest level of variation accounted for by environmental
factors alone was the late-successional, deciduous forest (19 to 50%). This was
hypothesized to be due to the selection of a microbial community, through
differences in carbon and nitrogen inputs and land management, which is more

responsive to changes in environmental parameters.

Chapter 3:

1.

There was a significant effect of agricultural treatment on evenness (P = 0.0476),
with the conventional agriculture treatment at 0-7 cm having the highest evenness.
This, along with rank/abundance curves, indicates that there are nirK OTU’s in
the never tilled and historically tilled treatments that may play an important role
in minimizing the amount of N20 emitted from sites currently not used for
agriculture.

Productivity (%C) did not have a significant relationship with measures of
denitrifier diversity, but an analysis using Spearman’s Correlation Coefficient
indicated a weak, positive relationship between relative disturbance and diversity.

Denitrifier communities show long-term effects from agriculture, as the 0-7 cm
and 13-20 cm communities from a site abandoned from agriculture 15 years ago
were still more similar to those from a site currently used for agriculture than to
those from a site never in agricultural use.

Communities in the 0-7 cm portion of the sites abandoned fi'om agriculture and

never tilled had diverged from those found at 13-20 cm, indicating that selective

125

pressures differ at each depth, most likely due to the differential level of influence
by plants at both depths.

A significant, positive relationship between rrn operon copy number and growth
rate of dissimilatory nitrate reducers was found, indicating that denitrifiers with

low and high rrn operon capy numbers differ in their ecological strategies.

Chapter 4:

1.

Three 16S rDNA sequences from libraries created with planctomycete-specific
PCR primers were similar to sequences from a group of planctomycetes capable
of anaerobic ammonia oxidation, but not enough to be affiliated with this group.
Due to the low coverage of the libraries, further study is warranted to determine
whether anammox is occurring or anammox capable planctomycetes are present
in soil.

Comparative analyses indicated that two soil treatments that have been used for
agriculture are more similar in planctomycetes community composition and
diversity measures to each other than to a treatment never used for agriculture. A
statistical comparison of the three libraries however, indicated no significant
differences in communities, suggesting that Planctomycetales are not contributing
to differences seen in total microbial community composition, may Only be
marginally affected by agricultural practices, and/or be capable of faster recovery

after cultivation than other microbial groups.

126

 

DIRECTIONS FOR FURTHER STUDY:

During the course of the work completed for this dissertation, additional
intriguing avenues for later study became apparent. More work is needed to identify the
denitrifying organisms whose nirK sequences dominate the KBS soil treatments never
used for agriculture and those abandoned from agriculture. Isolation of representatives of
these groups in pure culture would allow for physiological studies to determine whether
these organisms produce less N20 independent of the physical conditions under which
they are grown. The question of whether the dominant denitrifiers in the two treatments
are responsible for the lower N20 flux noted in these soil treatments as compared to the
flux from conventional agricultural soils could then be answered.

Another interesting avenue of research involving denitrifiers would be an
investigation into the significance of those organisms possessing one type of nitrite
reductase over the other. There are two known nitrite reductase enzymes, nirK and nirS,
which are mutually exclusive. These enzymes differ in their structure as well as their Km
for nitrite, with the average Km for nirK being higher than that of nirS [1]. This could
lead to situations in which denitrifiers with one type of Nir gene would be at a selective
advantage over those with the other gene in different environments. In molecular
surveys, it is not uncommon to find evidence for only one type of nitrite reductase (Table
5.1), as was the case for the experiments presented in Chapter 3. It is difficult to discern
a clear trend as to which Nir genes are found where, but in general, nirK organisms tend
to be found in a more diverse set of environments. Organisms possessing nirS are not
found in soil very often and seem to favor consistently wet environments, such as

activated sludge or water-body sediments. The lack of nirS-bearing organisms in soil

127

 

 

mm a .8 <2 .8. 80000: 888.88 0880 4020-08888 .880an ._0.88_880_.8w& 0:00 00m

:00 883—88050

R: + - 0:0 0882888000 b88300 80¢ 8088-8008800 wag-05888080 mew—00300 8000588080204

5 a + + 03030888893 02088885888 8883888: 05 28:2

no a + + 88830888 58808880000 883082 88.83003 05 :0 08800 088308.000 0O 88 80.8.0 8880::va

808 - + 8800 88.008 8880

+ + :00 8.0802

E - + :00 088—0: 0800.80.88

at e + e + 8880800298 .000 000000.83 0880:8803. go 0500 003M

.8. - 0008850 58 800008

HE + - ES 80008. a 50.: 03088—0 08950.0.

+ - 08808808 vesom 80w?—

W .E + + 8:08:08 885808 88:25:00 sewer—0.03

800 8080 85 + 8000 :80 + 88808 00 888888 .0858... Bazaar.

+ + 000883 8—3 80¢ 830.3

+ + com 8280—88 80888058 8.83 8.80.8.0 0880:8803

+ + com 80:05 82880—208 3.3 808.0 .8803

+ + 03088—0 80302030

HE + + «8.8803 02880-003508 w£©w£8800 00.0 0.8330 000.088.08.888”.—

8020038 3+0 3+0 8588888882808
388000.808 m0: 000000.888 v8.0.8:

 

m0: 0880 v8.0.8: 00.0 0.03.85 88=0§§>0m 086 038-

128

.80.“ 08000 0088 .<2 .8

.80080080 000 00000888 600 003 m0: 60000-8980 0088 085 088000.888 003 v2.8.8: .

 

 

:43 e + a + 308000.896 050 388000-85 Sufi-38¢ 00 800888-0880 080000-8088 =00 08.0—000.5
RE + + 80:88.00 8.08m
Had + + 08803 088258.80 0088033003 3928888888 0 0.80.8.0 new—088.0 020>30<
:3 + <2 00008.8 :83 war-8880888000 0 Eon-.8 owes—0 080>50<
83 - + 2880800898 000 888000.83 0880.80 088%»: 8.00.8.0 BEES 20:08.02”—
:0: + <2 BED .00 008000 05 to 8888-8200 8303 .80 088080 0088888888.: cowg

- + €088 $88305 0303 00008.00

- + 88800000 $03083 .0888.-

+ + 28800000 wfiBEw 0.80.008 880002
E 3 82888808.: 300 80 BEE-80.0 80.82880: .5580 $882398 :00 .EB_=o_-8w<
E. a + + 08883 9898888 80¢ :00 2083035—
3: + <2 00m 005$ 880000 05 .00 88:88:00 8803 warn-05309
+ + 8880.8. 8880:0090 88038003 80.8.0 new—020 020>to<
+ + 880.0 :25 Bob 0.88800 0802
mm a + + 00:0 025 80¢ =8 oBE<
b00833
:0: <2 .8. 808808802 0358 0:00 0880 08883888802 308m 800.80 80.0-c 08825000 E80050 058 =0m
3880.8003— T+V 34$ 0880:8880-8828m

008880.808 mam: 888000.888 v8.8.8:

 

”8880088080 8.0 2008.

129

may be due to a lesser ability to compete in this environment, but could also be due to
poor primer design [15] or seasonal effects [18]. Resolution of this issue would serve to
increase our knowledge regarding soil denitrifier ecology.

The last few suggested areas of interest deal with processes in the nitrogen cycle
besides denitrification by denitrifying microorganisms. Dissirnilatory nitrate reduction to
ammonium (DNRA) and denitrification by nitrifiers (nitrifier-denitrification) are two
microbial processes known to take place in soil that should receive more attention.
Under anaerobic conditions, DNRA organisms compete with denitrifiers for nitrate, but
unlike denitrification, under acidic conditions the end product of DNRA is not a gas that
is lost from the system. It has been hypothesized that DNRA organisms are selected for
in high carbon, low electron-acceptor (NO;°) environments and that the opposite
conditions select for denitrifiers [25, 26]. Therefore, the relative importance of each of
these dissimilatory processes may vary depending on the soil environment. For example,
at the KBS LTER the conventional agriculture treatment has a relatively high amount of
NO3' (6.54 :t 0.53 ug NO3-N g") and low amount of organic C (0.94 :h 0.05 kg C m'z)
while the opposite is true of the never-tilled mid-successional treatment (0.47 :t 0.03 ug
NO3-N g’I and 2.84 d: 0.22 kg C m'z) [27]. It would be worthwhile to determine whether
DNRA dominates in the never-tilled treatment and denitrification dominates in the
conventional agriculture treatment, as this may be a contributing factor to the low N20
flux from the never-tilled treatment.

In addition to determining the role of DNRA in the soil nitrogen cycle, a better
understanding is needed of the contribution of nitrifier-denitrification to nitrogen loss

fi'om soil. Ammonia oxidizing nitrifiers will denitrify when under stressful conditions,

130

such as low pH and low oxygen concentration. The end product of nitrifier-
denitrification can be N20 or N2, which contributes to nitrogen loss. The contribution of
nitrifier-denitrification to soil N20 production is still under debate, with present estimates
ranging from insignificant amounts [28] to ~30% of N20 production [29]. Owing to the
influence of environmental factors on whether nitrifier-denitrification occurs and the
difference in physical characteristics of soil at different sites, this process may vary in
relative importance between treatments, just as with DNRA.

The last suggestion for fmther study involves a recent discovery that would have
interesting repercussions in an agricultural soil environment. This was the observation of
anaerobic ammonia oxidation (anammox) occurring in a biofilrn from a denitrifying
wastewater treatment plant. 16S rDNA analysis indicated that the organism responsible
for anammox was a member of the Planctomycetales [30]. The net anammox reaction is
written as the equation: NI-Ia+ + N02' —> N2 + 2H2O and has a favorable Gibb’s free
energy change of AG°’ = —297 kJ/mol NH: [31]. The occurrence of this reaction in
wastewater treatment plants has proved effective in removing nitrogen sources as
dinitrogen gas [32]. In an agricultural setting, chemical fertilizers high in nitrogen
content are applied, and in most soils there exist anaerobic microsites within soil
aggregates [33]. These sites should be capable of supporting anaerobic microorganisms
that may be capable of using the anammox reaction as an energy supply. Theoretically,
anammox can convert NI-In+ added as chemical fertilizer and N02' from nitrification or
denitrification into dinitrogen gas. This N2 can then diffuse out of the soil, wasting a

significant amount of time, effort, and money ‘on the part of the farmer. Therefore,

131

determination of whether, and to what extent, anammox occurs in soil would be

important so that methods to control anammox could be developed.

132

 

REFERENCES:

1.

Zumft, W., Cell biology and molecular basis of denitrification. Microbiol. Mol.
Biol. Rev., 1997. 61(4): p. 533-616.

Braker, G., A. Fesefeldt, and K.-P. Witzel, Development of PCR primer systems
for amplication of nitrite reductase genes (nirK and nirS) to detect denitrifying

bacteria in environmental samples. Applied and Environmental Microbiology,
1998. 64(10): p. 3769-3775.

Hallin, S. and P.-E. Lindgren, PCR detection of genes encoding nitrite reductase
in denitrifiring bacteria. Applied and Environmental Microbiology, 1999. 65(4):
p. 1652-1657.

Braker, G., J. Zhou, L. Wu, A.H. Devol, and J .M. Tiedje, Nitrite reductase genes
(nirK and nirS) as fimctional markers to investigate diversity of denitrifying
bacteria in Pacific Northwest marine sediment communities. Applied and
Environmental Microbiology, 2000. 66(5): p. 2096-2104.

Braker, G., H.L. Ayala-Del-Rio, A.H. Devol, A. Fesefeldt, and J .M. Tiedje,
Community structure of denitrrfiers, Bacteria, and Archaea along redox gradients
in Pacific Northwest marine sediments by Terminal Restriction Fragment Length

Polymorphism analysis of amplified nitrite reductase (nirS) and 16S rRNA genes.
Applied and Environmental Microbiology, 2001. 67(4): p. 1893-1901.

Neufeld, J.D., B.T. Driscoll, R. Knowles, and ES. Archibald, Quantifying
functional gene populations: comparing gene abundance and corresponding
enzymatic activity using denitrification and nitrogen fixation in pulp and paper
mill effluent treatment systems. Canadian Journal of Microbiology, 2001. 47(10):
p. 925-934.

Nogales, B., K.N. Timmis, D.B. Nedwell, and AM. Osborn, Detection and
diversity of expressed denitrification genes in estuarine sediments after reverse
transcription-PCR amplification fiom mRNA. Applied and Environmental
Microbiology, 2002. 68(10): p. 5017-5025.

Prieme, A., G. Braker, and J .M. Tiedje, Diversity of nitrite reductase (nirK and
nirS) gene fragments in forested upland and wetland soils. Applied and
Environmental Microbiology, 2002. 68(4): p. 1893-1900.

Avrahami, S., R. Conrad, and G. Braker, Eflect of soil ammonium concentration
on N20 release and on the community structure of ammonia oxidizers and
denitrifiers. Applied and Environmental Microbiology, 2002. 68(11): p. 5685-
5692.

133

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

Liu, X., S.M. Tiquia, G. Holguin, L. Wu, S.C. Nold, A.H. Devol, K. Luo, A.V.
Palumbo, J .M. Tiedje, and J. Zhou, Molecular diversity of denitrifying genes in
continental margin sediments within the oxygen-deficient zone of the Pacific
coast of Mexico. Applied and Environmental Microbiology, 2003. 69(6): p. 3549-
3560.

Yan, T., M.W. Fields, L. Wu, Y. Zu, J.M. Tiedje, and J. Zhou, Molecular
diversity and characterization of nitrite reductase gene fragments (nirK and nirS)

from nitrate- and uranium-contaminated groundwater. Environmental
Microbiology, 2003. 5(1): p. 13-24.

Song, B.K. and BB. Ward, Nitrite reductase genes in halobenzoate degrading
denitrifying bacteria. Ferns Microbiology Ecology, 2003. 43(3): p. 349-357.

Henry, S., E. Baudoin, J .C. Lopez-Gutierrez, F. Martin-Laurent, A. Baumann, and
L. Philippot, Quantification of denitrifiring bacteria in soils by nirK gene targeted
real-time PCR. Journal of Microbiological Methods, 2004. 59(3): p. 327-335.

Qiu, X.Y., R.A. Hurt, L.Y. Wu, CH. Chen, J .M. Tiedje, and Z. Zhou, Detection
and quantification of copper-denitrifying bacteria by quantitative competitive
PCR. Journal of Microbiological Methods, 2004. 59(2): p. 199-210.

Throback, I.N., K. Enwall, A. Jarvis, and S. Hallin, Reassessing PCR primers
targeting nirS, nirK and nosZ genes for community surveys of denitrijying
bacteria with DGGE. Fems Microbiology Ecology, 2004. 49(3): p. 401-417.

J ayakumar, D.A., C.A. Francis, S.W.A. Naqvi, and BB. Ward, Diversity of nitrite
reductase genes (nirS) in the denitrifying water column of the coastal Arabian
Sea. Aquatic Microbial Ecology, 2004. 34(1): p. 69-78.

Wang, G. and H.D. Skipper, Identification of denitrifying rhizobacteria from

bentgrass and bermudagrass golf greens. Journal of Applied Microbiology, 2004.
97(4): p. 827-837.

Wolsing, M. and A. Prieme, Observation of high seasonal variation in community
structure of denitrifying bacteria in arable soil receiving artificial fertilizer and
cattle manure by determining T -RFLP of nir gene fragments. F ems Microbiology
Ecology, 2004. 48(2): p. 261-271.

Castro-Gonzalez, M., G. Braker, L. Farias, and O. Ulloa, Communities of nirS-
type denitrrfiers in the water column of the oxygen minimum zone in the eastern
South Pacific. Environmental Microbiology, 2005. 7(9): p. 1298-1306.

Sharma, 8., M.K. Aneja, J. Mayer, J .C. Munch, and M. Schloter, Characterization

of bacterial community structure in rhizosphere soil of grain legumes. Microbial
Ecology, 2005. 49(3): p. 407-415.

134

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Tsuneda, 8., R. Miyauchi, T. Ohno, and A. Hirata, Characterization of
denitrrfiring polyphosphate-accumulating organisms in activated sludge based on

nitrite reductase gene. Journal of Bioscience and Bioengineering, 2005. 99(4): p.
403-407.

You, S.J., Identification of denitrifying bacteria diversity in an activated sludge
system by using nitrite reductase genes. Biotechnology Letters, 2005. 27(19): p.
1477-1482.

Santoro, A.E., A.B. Boehm, and CA. Francis, Denitrifier community composition

along a nitrate and salinity gradient in a coastal aquifer. Applied and
Environmental Microbiology, 2006. 72(3): p. 2102-2109.

Sharma, S., Z. Szele, R. Schilling, J .C. Munch, and M. Schloter, Influence of
fi-eeze-thaw stress on the structure and function of microbial communities and

denitrifying populations in soil. Applied and Environmental Microbiology, 2006.
72(3): p. 2148-2154.

Tiedje, J .M., A]. Sexstone, D.D. Myrold, and J.A. Robinson, Denitrification -
Ecological Niches, Competition and Survival. Antonie Van Leeuwenhoek Journal
of Microbiology, 1982. 48(6): p. 569-583.

Tiedje, J .M., Ecology of denitrification and dissimilatory nitrate reduction to
ammonium, in Biology of Anaerobic Organisms, R. Mitchell, Editor. 1988, John
Wiley & Sons: New York. p. 179-244.

Robertson, G.P., B.A. Paul, and RR. Harwood, Greenhouse gases in intensive
agriculture: contributions of individual gases to the radiative forcing of the
atmosphere. Science, 2000. 289: p. 1922-1925.

Robertson, G.P. and J .M. Tiedje, Nitrous-Oxide Sources in Aerobic Soils -
Nitrification, Denitrification and Other Biological Processes. Soil Biology &
Biochemistry, 1987. 19(2): p. 187-193.

Webster, B.A. and D.W. Hopkins, Contributions from dififerent microbial
processes to N20 emission from soil under dtfi'erent moisture regimes. Biology
and Fertility of Soils, 1996. 22(4): p. 331-335.

Strous, M., J .A. Fuerst, E.H.M. Kramer, S. Logemann, G. Muyzer, K.T. van de
Pas-Schoonen, R. Webb, J.G. Kuenen, and SM. Jetten, Missing Lithotmph
Identified as New Planctomycete. Nature, 1999. 400: p. 446-449.

Mulder, A., AA. van de Graaf, L.A. Robertson, and LG. Kuenen, Anaerobic

Ammonium Oxidation Discovered in a Denitrifying F luidized Bed Reactor. FEMS
Microbiology Ecology, 1995. 16: p. 177-184.

135

 

32.

33.

Kuai, L. and W. Verstraete, Ammonium removal by the oxygen-limited
autotrophic nitrification-denitrification system. Applied and Environmental
Microbiology, 1998. 64(11): p. 4500-4506.

Tiedje, J.M., A.J. Sexstone, T.B. Parkin, N.P. Revsbech, and DR. Shelton,
Anaerobic Processes in Soil. Plant and Soil, 1984. 76(1-3): p. 197-212.

136

 

 

APPENDIX

A STRATEGY FOR CAPTURING PHYLOGENETICALLY USEFUL
INFORMATION IN BAC OR F OSMID LIBRARIES

INTRODUCTION :

Current estimates of bacterial numbers in soil are approximately 1 x 109 cells per
gram dry weight of soil, comprising greater than 105 species [1]. Cultivation attempts
routinely recover ca. 1% of this soil bacterial population [2, 3], making it difficult to learn
about the metabolic capabilities of not-yet-cultured organisms that may play important
ecological roles in the soil environment. As a result, studies employing molecular
techniques in place of cultivation have become common. These techniques allow an
assessment of the diversity of populations based on phylogenetically conserved genes
such as 16S rDNA to be performed. Unfortunately, these surveys do not provide
information about the metabolic potential of the organisms studied.

Recently, metagenorrric experiments employing the use of bacterial artificial
chromosome (BAC) or fosmid vectors have been successful in cloning large pieces of
environmental DNA [4, 5]. Typically, BAC and fosmid libraries have average insert
sizes of 50 to 100 kb and 40 kb respectively, so that when a clone containing a
phylogenetically conserved gene is identified, there are many other functional genes
captured as well. Sequencing of the clone then provides information not only about the
probable identity of the organism, but also about its metabolic capabilities.

One disadvantage of the classical BAC and fosrrrid library cloning strategies is
that extensive screening is required to identify the few clones containing phylogenetically

useful genes. For example, in a BAC library from soil made up of 24,400 clones, only 20

137

 

(ca. 0.1%) were found to possess rRNA genes [6]. A second disadvantage is that
screening of clones for rRNA genes with universal primers or probes is made more
difficult due to the presence of host cell DNA. Therefore, the strategy described in this
appendix was initiated in an effort to create libraries where all DNA cloned would have
rrn operon-bearing ends and hence be in a location on the BAC where end sequencing of
the insert would result in easy identification of its source. While the method ultimately
was not successfully employed using environmental DNA from soil, progress was made
in cloning DNA from a pure culture of Xanthomonas campestris pv. campestris ATCC

33913. Insights and progress made duringthe course of the project are reported.

DESCRIPTION OF BAC LIBRARY STRATEGY:

In order to capture rRNA genes in each BAC clone, a vector (SuperPhyloFOS)
was created containing a restriction site for I-CeuI (New England Biolabs, Beverly, MA).
I-CeuI is a homing endonuclease isolated from a large subunit rRNA gene of a
chloroplast in Chlamydomonas eugametos [7]. This endonuclease has been demonstrated
to recognize and cleave within a recognition site of approximately 26 bp present in most
prokaryotic 23S rRNA genes [8]. Cloning DNA that has been cut on one end with I-CeuI
will ensure that a portion of the 23S rRNA gene is inserted into the vector along with
other genetic information. Also, due to the conserved order of genes within the rRNA
operon, there is a high likelihood that the 16S rRNA gene would be captured along with
about one-third of the 23S rRNA gene (Figure 1).

The BAC cloning scheme successfully used to clone X. campestris DNA is

outlined in Figure 2 (see Figure 3 for a map of SuperPhyloFOS). The vector was

138

 

 

 

 

168 238 SS
I-CeuI
(26 bp site)

Figure 1: Typical orientation of an rRNA operon. Due to the nonpalindromic nature of
the I-CeuI cut site and the orientation of the site in pSuperPhyloBAC, approximately two-
thirds of the 23S rRNA gene and the complete 16S rRNA gene (arrow) of insert DNA

will be ligated into the vector.

139

Vector Genomic DNA

SuperPhyloFOS Xanthomonas campestris
i cut with Pmll SAP &

i cut with l-Ceul

...-airs
[sap

—
lcut with l- Ceul

size selection

fi\

Ligate
—.. --- m- ---8?
size selection, then
T4 Polynucleotide Kinase &

® T4 DNA Polymerase

l Ligate

_ 8“

. .
I ~‘
' D

Figure 2: Scheme followed during cloning of Xanthomonas campestris pv. campestris
ATCC 33913 DNA into SuperPhyloF OS. A (?) denotes a DNA end resulting from

shearing; therefore it is unknown whether the end has an overhang or is blunt.

140

Pmll

lacZa
\018
00%
P
0“;
SuperPhyloFOS

m 9,211 bp
‘8

o.-

A.“
For
repE

l-Ceu l

“'3

4.00

Figure 3: Map of the vector pSuperPhyloFOS (not drawn to scale). Section in light gray
is the insert from pSCANS. Black section is the pCClFOS backbone.

141

prepared by restricting 10.7 ug once with 40 U of Pmll (New England Biolabs) at 37°C
overnight and inactivating the Pmll at 65°C for 20 min. Pmll produces blunt ends which
were dephosphorylated by incubating at 37°C for 1 hr with 500 U of Shrimp Alkaline
Phosphatase (SAP; Roche, Indianapolis, IN) in the provided buffer and inactivating the
SAP at 65°C for 15 min. Next, the vector was cut once with 20 U of I-CeuI (New
England Biolabs) in NEBuffer 4 for 5 hrs at 37°C and then the I-CeuI was inactivated at
65°C for 20 min. This resulted in the vector being cut into two pieces; the desired
backbone into which DNA would be cloned and a small piece located between the Pmll
and I-CeuI cut sites whose removal resulted in the loss of lacZ expression. Size selection
was performed with a CHEF -DRTM H Pulsed Field Gel Electrophoresis (PFGE)
Apparatus (BioRad, Hercules, CA) set to 12°C, 100 V, and 1 to 2 s pulses for 15 hrs to
separate the two pieces of vector. The desired piece of vector was cut from the gel,
electroeluted into dialysis tubing containing 1 mL of 0.5X TBE, and concentrated.
Electroelution was performed using the PFGE apparatus set to 12°C, 50 V, and 1 to 2 s
pulses for 3.5 hours. A Centricon-IOO (Amicon, Inc., Beverly, MA) was used to
concentrate the electroeluted DNA using the manufacturer’s protocol. 20 ug of insert
DNA (X. campestris) was prepared by first dephosphorylating with 60 U of SAP in
NEBuffer 4 at 37°C for 1 hr and inactivating the SAP at 65°C for 15 min. The insert
DNA was then restricted by adding 60 U I-CeuI to the inactivated SAP reaction and
incubating for 5 hrs at 37°C followed by inactivation of I-CeuI at 65°C for 20 min. This
treatment of both vector and insert DNA resulted in only the I-CeuI cut ends possessing
phosphate groups; therefore these were the only ends capable of ligation. Since the I-

Ceul site is non-palindromic, the vector was unable to ligate to copies of itself. Next,

142

ligation with 2000 U T4 DNA ligase (New England Biolabs) was performed with a mix
of vector and insert DNA at 16°C overnight followed by inactivation of the ligase at 65°C
for 20 min. Size selection with a horizontal gel was performed with a 0.7% agarose gel
in sterile 0.5X TBE run at 4°C and 115 V for 1.5 hrs. This was done to eliminate small
pieces of DNA. DNA greater than ca. 23 kb in size was cut out of the gel, electroeluted
into 500 uL of 0.5X TBE at 115 V for 2 hrs, and the DNA ends blunted and
phosphorylated with an End-It” DNA-Repair Kit (Epicentre, Madison, WI) following
the manufacturer’s protocol. After this, another ligation reaction with the linear DNA
was set up using 800 U of T4 DNA ligase and incubation at 16°C overnight. This caused
the DNA to ligate into circular forms for transformation into E. coli by electroporation.
Screening of the library was performed by cutting the cloned BACs once, running them
on a horizontal gel and selecting those running at greater than the size of the BAC alone
for sequencing. The forward sequencing primer used for end sequencing of the inserts
was the pCClFOS forward primer (FP) provided with Epicentre’s CopyControl'm
F osmid Library Production Kit. A reverse sequencing primer (RP) was developed in the
lab that was specific to SuperPhyloF OS (5’ - GGT TGT AAC ACT GGC GAG - 3’).

The cloning scheme outlined above was arrived at through many experiments and
involved much trial and error. The three most challenging steps in creating a library
using the I-Ceul strategy were the creation of the vector, preparation of large pieces of
insert DNA and vector, and the Optimization of restriction, ligation and transformation.
These steps are commented on in more detail in the following sections in order to prOvide

guidance to others who may wish to use this or other large insert library strategies.

143

CONSTRUCTION OF THE BAC VECTOR:

A new vector, SuperPhyloF OS, was constructed which contains an I-Ceul
restriction site (Figure 3). The vector was constructed so that it could be used as a BAC
or fosmid, although it was only employed as a BAC in the experiments described here.
SuperPhyloFOS was created by inserting a portion of the pSCANS plasmid (a gift fi'om
Dr. John Dunn of Brookhaven National Laboratory) containing an I-CeuI site and
kanamycin resistance gene into a pCClFOS (Epicentre) plasmid. When SuperPhyloFOS
is transformed into E. coli strain JW366 (sold as One ShotTM Electrocompetent
GeneHogs from Invitrogen; Carlsbad, CA or EPI3001'M-TlR Phage Tl-Resistant E.coli
from Epicentre), it is inducible to high copy number upon addition of L-arabinose to the
media. E. coli JW366 contains a defective transcription factor (TrfA) linked to an
arabinose promoter that will act on the oriV of the vector. Replication begins and since
the defective transcription factor cannot disengage from the oriV, replication continues
indefinitely. For blue/white selection to occur, the vector needs to be cut with both I-
Ceul and a restriction enzyme with a site within the lacZo. gene. Both kanamycin and
chloramphenicol resistance genes are present for antibiotic selection.

The identity and properties of SuperPhyloF OS were confirmed experimentally.
The sequences of both pSCANS and pCClF OS are known, and using this information, a
map of the theoretical sequence of SuperPhyloF OS was made. Restriction digests were
used to confirm that the desired cloning sites were present and that the I-CeuI site could
be restricted properly. The lacZa gene was inducible with IPTG and cells were blue
when in the presence of X-gal. It is important to note that induction with IPTG is

required for colonies without inserts to turn blue on selective plates. Induction with L-

144

arabinose to high copy number was also functional as seen in comparisons of plasmid
preparations from induced and non-induced cells. Once the identity and desired
properties of SuperPhyloFOS were confirmed, attempts to clone DNA from a pure

culture of X. campestris were made to test whether the cloning scheme was feasible.

PURE CULTURE AND ENVIRONMENTAL DNA EXTRACTION:

Pure culture DNA: Embedding cells within agarose plugs prior to cell lysis and
DNA restriction is commonly used in construction of large insert libraries in order to
reduce the amount of DNA shearing. However, since lysis and DNA restriction within
plugs was found to be inconsistent between different batches of plugs the Marmur
procedure [9, 10] was used to isolate genomic DNA. Genomic DNA fi'om a pure culture
of X. campestris isolated using the Marmur procedure yielded DNA ranging in size from
24 to 145.5 kb.

Environmental DNA: Initially, indirect DNA extraction methods were employed
to obtain DNA from soil cells. Indirect methods were used first as they are less likely to
result in the shearing of DNA because cells are extracted from the soil matrix before
lysis. Blending, sonication, and combinations of the two were applied to soil in an effort
to dislodge cells from sediment particles and isolate them from the soil matrix. The
supernatant created fi'om these techniques was then applied to an Optiprep density
gradient (Axis-Shield, Oslo, Norway) with a density of 1.320 g mL". Extracted cells
were cast into plugs of 1% InCert agarose (Carnbrex Bio Science Rockland, Inc.,
Rockland, ME), stored until lysed, and after lysis the DNA was sized on a PFG. The

indirect extraction method was inefficient, resulting in percent recoveries of cells ranging

145

from 0.09 to 1.9%. Most cells were lost during the extraction process when large
particles were allowed to settle out of the supernatant, resulting in the settling out of any
cells still attached to particulate. An additional issue was the presence of hurrric acids
along with cells, which interfered with subsequent cell lysis and restriction within the
agarose plugs.

Due to the above mentioned issues, a direct method of DNA extraction from soil
based on that of Zhou and colleagues [11] was developed. This method has been shown
to produce minimal shearing of DNA and employs a buffer containing sodium dodecyl
sulfate (SDS), hexadecyltrimethylarnmonium bromide (CT AB), and proteinase K. A
chloroform-isopropanol precipitation is then performed to recover the nucleic acids. As
with the indirect extraction method, humic acid contamination is routine. To reduce the
amount of humic acids present in the DNA, extracted samples were run on a horizontal
gel. Large pieces of DNA moved slowly as one large band through the gel, while hurnics
were drawn through more quickly. After the section of gel containing large pieces of
DNA was excised, electroelution in a PFG apparatus was used to extract the purified
DNA. DNA was then rinsed and concentrated using a Centricon-lOO and sized by
PFGE. DNA sizes ranged widely, fi'om less than 9.4 kb to greater than 97 kb (Figure 4),

which was considered acceptable for construction of BAC or fosmid libraries.

OPTIMIZATION OF RESTRICTION, LIGATION, AND TRANSFORMATION:
Restriction of vector and insert: Restriction of both vector and insert was found
to work best under conditions where contamination from cellular or environmental

components was minimized. For this reason, SuperPhyloFOS was isolated from the

146

12 3 4 5 6 7 8 91011121314151617

 

Figure 4: PFG of indirectly extracted soil DNA. Lanes 1 & 17 are Lambda Ladder PFG
Markers; 2 & 16 are Mid-Range H PFG Markers, and 3 & 15 are Low-Range PFG
Markers (all from New England Biolabs). Lanes 4 - 5 are fiom two extracts of deciduous
forest soil; Lanes 6 - 7 are from two extracts of mid-successional, never-tilled soil. Lanes
8 - 11 are from four extracts of soil abandoned from agriculture; and Lanes 12 - 14 are
fi'om three extracts from conventional agricultural soil.

147

genomic DNA of its host cell with a CsCl gradient and concentrated by ethanol
precipitation [12]. The CsCl gradient resulted in DNA with fewer contaminants and
DNA yields were higher than with standard plasmid preparation kits. Genomic DNA
from pure cultures was isolated with the Marmur procedure as described previously,
resulting in minimal contamination fi'om cellular components. As with pure cultures, the
restriction of DNA fi'om soil cells embedded in agarose plugs was problematic. There
was a high amount of variability in the ability to lyse cells within the plugs and restriction
required optimization for each batch of plugs. This was presumably due to the presence
of humic acids and other contaminants co-extracted with the DNA.

Ligation: Ligations with X. campestris DNA were set up to maximize the amount
of insert DNA present; therefore, molar ratios ranging from 1:10 to 1:50 of vector to
insert were commonly used (assuming an average X. campestris DNA size of 20 kb),
although preliminary results indicate that a 1:10 ratio worked best. The 1:10 ratio
ligation resulted in 3 out of 10 clones containing inserts, while the 1:50 ratio ligation
resulted in only 1 out of 10 clones containing inserts.

Transformation: A study was performed to determine what voltage would
maximize the transformation efficiency of clones with inserts of approximately 100 kb,
as well as whether desalting of the sample had an effect on transformation efficiency.
Two clones from a common bean (Phaseolus vulgaris L.) BAC library constructed by
Melotto and colleagues [13] were used in the study along with a 12 kb plasmid.
Electroporation was carried out with 50 ng of DNA in cuvettes with 1 mm gaps on a
GenePulser Xcell (BioRad) set to 100 Q and 25 uF. Desalting was performed with

agarose cones as recommended in Appendix B of the Epicentre CopyControl"'M BAC

148

Cloning Kit. The best voltage for transformation of both BACs was 1300 V (Figure 5),
while for the small plasmid it was 1700 V. Generally, as the size of the transformed
DNA increased, its overall transformation efficiency decreased. Desalting resulted in
increases in transformation efficiency of the 115 kb and 79 kb BACs of 37% and 142%,
respectively. Therefore, a transformation voltage of 1300 V and desalting of the ligation

reaction prior to transformation is recommended when creating BAC libraries.

SUMMARY:

The cloning scheme outlined in Figure 2 was successfully employed in cloning
genomic DNA fiom X. campestris. Of the 20 clones screened by end sequencing of the
insert, the largest insert size found was only 7.3 kb making the maximization of insert
size the next logical step in this particular set of experiments. Once optimized, the
procedure could be used to create BAC libraries of environmental DNA.

One option that was proven effective by Nesba and colleagues [14] was very
similar to that proposed here. An adaptor containing the I-Ceul recognition sequence was
ligated into the pCClFOS (Epicentre) vector and the new construct used as a fosmid
instead of a BAC. Fosmids are packaged as linear DNA into lambda phage particles and
then inserted into a host cell through transduction; the DNA is then circularized and
maintained in the cell. Because fosmid packaging requires a cos site (located on the
vector) and a total of approximately 40 kb of linear DNA, all clones will contain large
insert DNA ligated to the vector. SuperPhyloF OS also should be ftmctional as a fosmid,
and this may be the most efficient way to construct genomic libraries using the I-Ceul

strategy.

149

 

3’ r

:2 10 i

E:

‘5, 107-

5

E 1064 j

\1

III

3 105-

'8

E 104-

0

El 103-

r:

102 .

 

500 ' 1000 ' 1500 ' 20'00 ' 2st
Voltage

Figure 5: The dependence of transformation efficiency on voltage and desalting of
sample prior to transformation. (I) 12 kb plasmid control, (V) 79 kb BAC, (e) 115 kb
BAC, (V) desalted 79 kb BAC, and (o) desalted 115 kb BAC.

150

REFERENCES:

1.

10.

Gans, J ., M. Wolinsky, and J. Dunbar, Computational improvements reveal great
bacterial diversity and high metal toxicity in soil. Science, 2005. 309(5739): p.
1387-1390.

Staley, J .T. and A. Konopka, Measurement of Insitu Activities of

Nonphotosynthetic Microorganisms in Aquatic and Terrestrial Habitats. Annual
Review of Microbiology, 1985. 39: p. 321-346.

Amann, R., W. Ludwig, and K. Schleifer, Phylogenetic identification and in situ
detection of individual microbial cells without cultivation. Microbiol. Rev., 1995.
59(1): p. 143-169.

Beja, 0., MT Suzuki, E.V. Koonin, L. Aravind, A. Hadd, L.P. Nguyen, R.
Villacorta, M. Amjadi, C. Garrigues, S.B. Jovanovich, R.A. Feldman, and BF.
DeLong, Construction and analysis of bacterial artificial chromosome libraries

fi'om a marine microbial assemblage. Environmental Microbiology, 2000. 2(5): p.
5 16-529.

Rondon, M.R., P.R. August, A.D. Bettermann, S.F. Brady, T.H. Grossrnan, M.R.
Liles, K.A. Loiacono, B.A. Lynch, I.A. MacNeil, C. Minor, C.L. Tiong, M.
Gilrnan, M.S. Osbume, J. Clardy, J. Handelsman, and RM. Goodman, Cloning
the soil metagenome: a strategy for accessing the genetic and functional diversity
of uncultured microorganisms. Applied and Environmental Microbiology, 2000.
66(6): p. 2541-2547.

Liles, M.R., B.F. Manske, S.B. Bintrim, J. Handelsman, and RM. Goodman, A
census of rRNA genes and linked genomic sequences within a soil metagenomic
library. Applied and Environmental Microbiology, 2003. 69(5): p. 2684-2691.

Marshall, P. and C. Lemieux, Cleavage Pattern of the Homing Endonuclease
Encoded by the 5th Intron in the Chloroplast Large Subunit Ribosomal-Rna-
Encoding Gene of Chlamydomonas-Eugametos. Gene, 1991. 104(2): p. 241-245.

Liu, S.-L., A. Hessel, and KR. Sanderson, Genomic mapping with I-Ceu I, an
intron-encoded endonuclease specific for genes for ribosomal RNA, in Salmonella
spp., Escherichia coli, and other bacteria. Proc Natl Acad Sci, 1993. 90: p. 6874-
6878.

Marrnur, J ., Procedure for Isolation of Deoxyribonucleic Acid from Micro-
Organisms. Journal of Molecular Biology, 1961. 3(2): p. 208-&.

Johnson, J .L., Similarity Analysis of DNAs, in Methods for General and

Molecular Bacteriologr, P. Gerhardt, et al., Editors. 1994, American Society for
Microbiology: Washington, DC. p. 655-682.

151

11.

12.

l3.

l4.

Zhou, J .Z., M.A. Bruns, and J .M. Tiedje, DNA recovery from soils of diverse
composition. Applied and Environmental Microbiology, 1996. 62(2): p. 316-322.

Sarnbrook, J. and D.W. Russell, Molecular Cloning: A Laboratory Manual. 3rd
ed. 2001, Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press.

Melotto, M., M.F. Coelho, A. Pedrosa-Harand, J .D. Kelly, and L.E.A. Carnargo,
The anthracnose resistance locus Co-4 of common bean is located on
chromosome 3 and contains putative disease resistance-related genes. Theoretical

and Applied Genetics, 2004. 109(4): p. 690-699.
Nesbo, C.L., Y. Boucher, M. Dlutek, and W.F. Doolittle, Lateral gene transfer

and phylogenetic assignment of environmental fosmid clones. Environmental
Microbiology, 2005. 7(12): p. 201 1-2026.

152

   

1111111111111111111