,5... ‘ ‘
éfi .
.331? a.»

s a. «war,

i;

z! a

f . g!

'1 ‘12...

.5335 if
xiii:

x.“

, x5.

Vlalel‘nt

33.3???
5-:

. liar... .1
3.. .223...
1...: i... .1
i I???

. ,

 

.
2 ...

flu (v.4. .111: .
, , .2
a LA
‘ am... aw...
v 1,.\

 

l...Y ,.
. . 1 .

 

 

 

UBRARY

F “ L C+A+A
IVIIUI “UCIII Ulalc

___9_“5V9r§ilyl -_

— .- ‘oofi-vlv- - u-uv- ”w—‘

 

This is to certify that the
dissertation entitled

COMPARISONS OF METHANOTROPH COMMUNITIES IN SOILS
THAT CONSUME ATMOSPHERIC METHANE

presented by

Uri Yitzhak Levine

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Microbiology and Molecular
Genetics

 

 

\ /" / ‘

 

 

V w
' L%}Maj3’r Professor’s Signature
7 — 36 - 09
Date

MSU is an Alfinnative AcfiorVEquaI Opportunity Employer

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KLIProj/AccaPreleIRC/DaleDue.indd

 

COMPARISONS OF METHANOTROPH COMMUNITIES IN SOILS THAT
CONSUME ATMOSPHERIC METHANE

By

Uri Yitzhak Levine

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Microbiology and Molecular Genetics

2009

ABSTRACT

COMPARISONS OF METHANOTROPH COMMUNITIES IN SOILS THAT
CONSUME ATMOSPHERIC METHANE

By
Uri Yitzhak Levine

Methane is a potent greenhouse gas that is 21-25 times more efficient at trapping
heat (infrared radiation) than carbon dioxide. Methane oxidation is mediated by methane-
consuming microbes (methanotrophs), but only in upland soils does the activity of
aerobic methanotrophs account for a net uptake of atmospheric methane. The conversion
of native lands to row-crop agriculture diminishes the strength of the soil methane sink,
typically dropping the rate of methane consumption by 70%.

To determine the relationship between the rates of methane consumption in soils
and the diversity of microbes that catalyze them, we conducted molecular surveys of
methanotroph communities across a range of land uses at The Kellogg Biological Station
Long Term Ecological Research Site (KBS LTER) and correlated our findings to
measurements of the in situ fluxes of methane. Rates of methane consumption and
methanotroph diversity were positively correlated, as conversion of lands to row-crop
agriculture led to a 7-fold reduction fiom maximal rates of consumption in the native
deciduous forests. In fields abandoned from agriculture both methanotroph richness and
the consumption of methane were estimated to require approximately 75 years to return
to the present diversity and consumption rate of the native deciduous forests. The linear
trajectory for recovery of both measures suggested that managing lands to conserve or
restore methanotroph diversity would yield increases in the rate of consumption of

atmospheric methane in KBS LTER soils.

Long-term fertilization is one aspect of row-crop agriculture that is likely to be a
significant disturbance to the methanotroph community. We hypothesized that a
consequence of this disturbance would be its association with decreases in methane
consumption and methanotroph richness in fertilized forest sub-plots at KBS LTER, but
neither rate nor methanotroph richness declined due to long-term fertilization alone at
KBS LTER. A meta-analysis examined the effect of long-term fertilization in other sites,
and revealed no consistent decline in methanotroph richness in fertilized soils. The
methanotroph communities did display a distinct biogeography with communities
clustering together based on geographic location. As a consequence, the composition of
the unique soil methanotroph community probably plays a role in dictating the response
of the methanotroph community to changing land use and its disturbances.

The causes behind the change in methanotroph richness and the correlated
decrease in methane consumption associated with row-crop agriculture at KBS LTER
remains unclear as it is not caused by fertilization alone. The quantification of the effect
of other variables associated with agricultural management is necessary to determine
which management strategies at KBS LTER could be utilized to enhance methanotroph
richness. However, a management strategy determined at KBS LTER to be beneficial to
the methanotroph community may not be applicable to other paired sites, as the unique
methanotroph community and environmental characteristics from each geographic

location will probably yield a dissimilar response to the management practice.

ACKNOWLEDGEMENTS

There are many people who have helped me on my journey through graduate
school, and I am grateful to them all. My colleagues in the Schmidt Lab were
enormously generous with their time, advice and help. Without their lessons, advice and
training I would not be where I am today. Stephanie Eichorst took me under her wing as a
rotation student and helped teach me many of the bench skills I would repeatedly utilize
during my research. Zarraz Lee was the longest fellow traveler and her help, ideas and
encouragement were invaluable. In my first few years of graduate school, John Breznak,
John Wertz, Kristin Huizinga, Brad Stevenson, Jorge Rodrigues, Dion Antonoplulous and
Kwi Kim were all tremendous sources of advice and help, and later on so too were Clegg
Waldron, Tracy Teal, Vicente Gomez, Kevin Theis and Russell Grant. Just as important,
all members of the Schmidt lab established a comfortable and collegial working
environment that fostered my development, was focused on helping and supporting each
other, and is one that I hope to help recreate wherever I go so that I and others can
continue to reap its benefits.

That atmosphere was a product of my advisor, Tom Schmidt, who set the tone
through his own unfailing willingness to help. In addition to setting an excellent example
for me to follow, through his tutelage I have become a much better scientist who is more
meticulous, thoughtful, purposeful, and independent. The lessons and thinking that he
has instilled in me I will take with me throughout my career, and I am extremely grateful.
In addition, graduate school has been a period of tremendous growth for me in my
personal life, and I am thankful to have had an advisor who was understanding and

patient as I adjusted to becoming a husband and father.

iv

My committee, Phil Robertson, Rob Britton, Terry Marsh, and Ned Walker were
invaluable resources, each contributing expertise to different areas of my research,
expertly guiding my research, and they never failed to help or to share their own lab’s
resources in helping my research. The Britton lab graciously shared much equipment,
and without the help of Andrew Corbin, Stacey VanderWulp, Neville Millar and the rest
of the Robertson lab, my sampling and experiments at the Kellogg Biological Station
Long Term Ecological Research Site (KBS LTER) would have been impossible.

The Robertson lab was also responsible for the taking of all of trace gas fluxes
used in this study, and without their collaboration my data could not have been placed in
its proper context, and its potential significance would remain unclear. I am grateful for
their hard work, willingness to help, and the ease of our collaboration. When I traveled
to Rothamsted Research I was graciome hosted by the Hirsch lab, and Penny Hirsch, Ian
Clark, Keith Goulding and Paul Poulton all devoted considerable time and effort to
ensuring that my visit and research was productive.

All of this research would not have been possible without funding from an
Environmental Protection Agency Star Fellowship, College of Natural Science Recruiting
Fellowship and Dissertation Completion Fellowship, a Kellogg Biological Station Long
Term Ecological Research small graduate student grant, international travel support from
the National Science Foundation’s Long Term Ecological Research program, and support
from the Michigan Agricultural Experimental Station.

Last, but certainly not least, I would not be here without the support of my friends

and especially my family. My wife Rachel and daughter Channa have been my

enthusiastic cheerleaders, enduringly patient, and willing to support me in all of my

endeavors. I am and will always be, eternally gratefirl for them and their support.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... ix

LIST OF FIGURES ...............................................................................

CHAPTER 1. Uncertainties in the Global Methane Budget and Biological
Influences on the Concentration of Atmospheric Methane ..............................

....X

...1

Introduction .................................................................................... 1
Summary ...................................................................................... l4
Thesis Overview ................................................................................ 14

References ..................................................................................

CHAPTER 2. Agriculture’s Impact on Microbial Diversity and the Flux of

..20

Greenhouse Gases ................................................................................... 26

Abstract ......................................................................................
Introduction .................................................................................
Materials and Methods ...................................................................

Results
Drscussmn

Conclusion............
References ..................................................................................

CHAPTER 3. The Impact of Long-term Fertilization to the Methanotroph

..26
..27
..30
.37
.42

Acknowledgments..................................

.48
..63

Communities in Soils ................................................................................ 68

Abstract ......................................................................................
Introduction .................................................................................
Materials and Methods ...................................................................

Results
Discussion.............
Conclusron
Acknowledgments.............................

References ..................................................................................

CHAPTER 4. Conclusions and Future Directions.....................

..68
..69
..72
.77
.79
.85
.86
..97

.100

References ................................................................................... 105

APPENDIX A. Assessment of the PCR bias .................................................. 107

Introduction .................................................................................

107

Methods ..................................................................................... 108

Results
Discussronn

.110

111

References ................................................................................... 1 l6

vii

APPENDIX B. Ammonia and Nitrate Before and After Fertilization in a KBS

LTER Fertilized Sub-plot ......................................................................... 117
Introduction ................................................................................. 1 17
Methods ..................................................................................... 117
Results and D1scussron118
Acknowledgmentsll9
References ................................................................................... 121

APPENDIX C. Biogeography of Methanotrophs is Well Drained Soils... .. ............122
Introduction ................................................................................. 122
Methods ..................................................................................... 122
Results and Discussron123
References ................................................................................... 127

viii

LIST OF TABLES

Table 2.1. KBS-LTER Sites Investigated in this study ......................................... 49
Table 2.2. Summary of pmoA clone libraries ................................................... 50
Table 2.3. Correlations between the flux of greenhouse gases, species richness and

environmental conditions ........................................................................... 51
Table 3.1. Summary of the Sites and pmoA libraries used in this study. .................... 87

Table A. 1. Composition, expected output and actual output of defined artifical

communites of pmoA and amoA species used to assess the PCR bias of the primer pair
A189-A682. ........................................................................................ 114

Table C.l. Summary of the Sites and pmoA libraries used in this study. ................. 125

ix

LIST OF FIGURES

Figure 1.1. A simplified schematic of the metabolic pathway of methane oxidation. The
key enzyme is the methane monooxygenase, which facilitates the oxidation of methane
into methanol .......................................................................................... 18

Figure 1.2. Phylogenetic tree of selected partial pmoA and amoA protein sequences from
public databases. The tree is based on 164 amino acid positions using Phylip Protein
Maximum Likelihood as implemented in ARB (Ludwig et a1. 2004). Boxed and shaded
labels are indicative of pmoA clades that have been found in soils, have not been
cultured, and are thought to likely play a substantial role in the oxidation of atmospheric
methane. The scale bar represents 10 PAM units ............................................... 19

Figure 2.1. Average monthly carbon dioxide production and methane consumption based
on current land management and historical land use at the KBS-LTER: Agricultural
management of historically tilled land (Ag HT; V ), early successional plant communities
on fields that had been abandoned from agriculture in 1989 (Early HT; I ), mid-
successional plant communities on either historically tilled land (Mid HT; J- ) or never
tilled land (Mid NT; 0 ), or a late successional deciduous forest (Late DF; 0 ) ............ 52

Figure 2.2. The effect of different land uses on average carbon dioxide emission (a) and
net methane consumption (b) at KBS LTER. Different letters represent significant
differences (p<0.05) between treatments. Rate measures are the same as those in Fig. 2.1.
Error bars represent standard errors. Land use treatments are: agricultural management of
historically tilled land (Ag HT), early successional plant communities on fields that had
been abandoned from agriculture in 1989 (Early HT), mid-successional plant
communities on either historically tilled land (Mid HT) or never tilled land (Mid NT), or
a late successional deciduous forest (Late DF) ................................................. 53

Figure 2.3. pmoA rarefaction curves from all of the KBS-LTER pmoA clone libraries
constructed from the Late DF and Ag HT treatments. Libraries were constructed from
DNA extracted from soil sampled in December 2004, June 2005, and June 2006 with
various annealing temperatures (For additional details see Table 2.2, and Methods). All
curves were constructed using data from neighbor joining matrixes from Arb (Ludwig et
a1. 2004), with curves calculated by DOTUR (3 8). Methanotroph species are defined as
pmoA sequences having 94% average nucleotide sequence similarity. Error bars
representing 95% confidence intervals were omitted for the sake of c1arity...... . . . . . . . . . . .54

Figure 2.4. Phylogenetic tree of selected partial pmoA and amoA protein sequences from
public databases and translated from PCR-based clone libraries fiom KBS-LTER soils.
The tree is based on 164 amino acid positions using Phylip Protein Maximum Likelihood
as implemented in ARB (Ludwig et al. 2004). Boxed or circled labels are indicative of
pmoA clades recovered fi'om KBS-LTER soils. Cluster I and Cluster II clades (boxed
and starred labels) were recovered from Ag HT and Late DF, while clades KBSl, JR1 ,
Upland Soil Cluster or, and MRI (boxed labels) were recovered from Late DF. Clade

RA21 (oval) was only recovered from Mid NT soil. The scale bar represents 10 PAM
unit ..................................................................................................... 55

Figure 2.5. Neighbor joining phylogenetic tree of the partial nucleic acid sequences of
pmoA and amoA fi'om reference sequences, and KBS-LTER clone libraries from soil
sampled in December 2004, June 2005, June 2006, and June 2007. The tree is rooted
with the branching from the amoA sequences from Nitrosomonas eutropha and
Nitrosococcus mobilis. Nodes representing the 27 different methanotroph species
(defined as pmoA sequences having 94% average nucleotide sequence similarity)
identified at KBS-LTER are highlighted. The colors of the highlights reflect species’
membership in the seven pmoA clades depicted in Figure 5. The pmoA sequences
included in this tree are deposited in GenBank under accession numbers F] 529724 -
FJ529808, and GQ219582-GQ219583 ............................................................ 56

Figure 2.6. The correlation between methanotrOph diversity and methane consumption at
KBS LTER. Relationship between summer methane consumption (June-August) and
methanotroph richness (both represent averages of 3 replicate plots) across landscapes at
the KBS LTER. A simple linear regression is presented (r2 =0.62, p <0.001) with
operational taxonomic units (OTUs) defined as peaks in the tRF LP analysis that have
been identified as a pmoA gene. Symbols are as follows: Agricultural management of
historically tilled land (Ag HT; ‘7 ), early successional plant communities on fields that
had been abandoned from agriculture in 1989 (Early HT; I ), mid-successional plant
communities on either historically tilled land (Mid HT; 2 ) or never tilled land (Mid NT;
0 ), or a late successional deciduous forest (Late DF; 0 ) ..................................... 59

Figure 2.7. The recovery of methanotroph diversity and methane consumption at KBS
LTER following row-crop agriculture. Increase in methanotmph diversity (open
symbols) and methane consumption (closed symbols) as a fimction of time since
cesssation of agriculture. Measurements of the deciduous forest (Late DF) are positioned
based on projections fiom linear regression used to fit methanotroph diversity (y =0.07x
+2.05; r2=0.99, p = 0.020) or methane consumption (y=0.13x + 0.80; 1’2 =0.69, p < 0.001).
Error bars represent standard errors; symbols are as follows: Agricultural management of
historically tilled land (Ag HT; V ), early successional plant communities on fields that
had been abandoned from agriculture in 1989 (Early HT; I ), mid-successional plant
communities on either historically tilled land (Mid HT; :3. ) or never tilled land (Mid NT;
0 ), or a late successional deciduous forest (Late DF; 0 ) ..................................... 60

Figure 2.8. Similarity of methanotroph communities among KBS-LTER treatments.

The Serenson index was calculated for each pairwise comparison of methanotrophs using
OTUs for all confirmed pmoAs, and then plotted using two-dimensional non-metric
multidimensional scaling (Hammer et al. 2001). Symbols are as follows: Agricultural
management of historically tilled land (Ag HT; V ), early successional plant communities
on fields that had been abandoned from agriculture in 1989 (Early HT; I ), mid-
successional plant communities on either historically tilled land (Mid HT; .3. ) or never
tilled land (Mid NT; 0 ), or a late successional deciduous forest (Late DF; 0 ). Figure 2.9
contains the same data represented as a dendogram) ............................................. 61

xi

Figure 2.9. Similarity of methanotroph communities among KBS-LTER treatments. The
Serenson index was calculated for each pairwise comparison of methanotrophs using
OTUs for all confirmed pmoAs, and then clustered using neighbor-j oining with MEGA
(54). Symbols are as follows: Agricultural management of historically tilled land (Ag
HT; V ), early successional plant communities on fields that had been abandoned fiom
agriculture in 1989 (Early HT; I ), mid-successional plant communities on either
historically tilled land (Mid HT; .5». ) or never tilled land (Mid NT; 0 ), or a late
successional deciduous forest (Late DF; 0 ) ....................................................... 61

Figure 3.1. Differences between Late DF fertilized and control treatments at KBS LTER
in average net methane consumption and the average number of methanotroph species
(defined as pmoA sequences having 94% average nucleotide sequence similarity).
Fertilization has no effect on either measure (t=0.58, p=0.57 for consumption; and t=1.18,
p=0.36 for richness). Measures for net methane consumption are averages from 3
replicates of each treatment, while the methanotroph species are averages fi'om 2
replicates of each treatment. Net methane consumption from the same 2 replicates as
those with methanotroph species data yields the same result (20.9:l:4.3 and 19.2:l:3.3
average net methane consumption of the control and fertilized sub-plots, respectively,
and t=0.28 p=0.79). Error bars represent standard error ........................................ 88

Figure 3.2. (A) Rarefaction curves from pmoA clone libraries constructed from the
control and fertilized sub-plots of the late successional forest at KBS LTER. Libraries
were constructed from DNA extracted from soil sampled in June 2007 . (B) Rarefaction
curves fi'om pmoA clone libraries constructed fi'om Rothamsted Reasearch soils sampled
in Feburary 2008. All curves were constructed using data from neighbor joining matrixes
from Arb (Ludwig et al. 2004), with curves calculated by DOTUR (Schloss and
Handelsman 2005). Methanotroph species are defined as pmoA sequences having 94%
average nucleotide sequence similarity. Error bars representing 95% confidence intervals
were omitted for the sake of clarity ...................................... ' .......................... 89

Figure 3.3. Phylogenetic tree of selected partial pmoA and amoA protein sequences fi'om
public databases and translated from PCR-based clone libraries from late successional
deciduous forest sub-plots. The tree is based on 164 amino acid positions using Phylip
Protein Maximum Likelihood as implemented in ARB (26). Bolded clades were
recovered from the late successional forest sub-plots. The symbols adjacent to the clade
reflect the clade’s recovery from either the control treatment (0) or the fertilized
treatment (I). The scale bar represents 10 PAM units. ...................................... 90

Figure 3.4. (A) Similarity of methanotroph communities among fertilized and control
sub-plot replicates. (B) Similarity of methanotroph communities between fertilized and
control sub-plot replicates with all other pmoA libraries from KBS LTER. Expect for the
fertilized and control sub-plots, the pmoA clone libraries were first reported in Chapter 2,
and details of their construction can be found there. Both dendograms are based on
Serenson index calculations for each pairwise comparison of the methanotroph
communities using pmoA species (defined as pmoA sequences having 94% average
nucleotide sequence similarity), and then clustered using neighbor-joining with MEGA
(35). The scale bars represent at 0.05 change in the Sarenson index. ........................ 91

xii

Figure 3.5. Neighbor joining phylogenetic tree of the partial nucleic acid sequences of
pmoA and amoA fi'om reference sequences, and KBS-LTER clone libraries from the
control and fertilized sub—plots soil sampled in June 2007. The tree is rooted with the
branching from the amoA sequences from Nitrosomonas eutropha, Nitrosococcus
mobilis, and Nitrosovibrio tenuis. Nodes representing the 17 different methanotroph
species (defined as pmoA sequences having 94% average nucleotide sequence similarity)
found in late successional deciduous forest sub-plots are circled. The symbols in the
circles reflect species’ recovery from either the control treatment (O ) or the fertilized
treatment (I). ........................................................................................ 92

Figure 3.6. Methanotroph richness in paired sites featuring long-term fertilized (black
shading) and non-fertilized soils (grey shading). The three sites to the left compare
fertilized sites in the context of agricultural management, while the two sites on the right
are fertilized forests whose only land management is fertilization. Error bars represent
standard errors as the averages are reported for sites that featured replicated sites.
Additional site details are provided in Table 3.1 .................................................. 94

Figure 3.7. Phylogenetic tree of selected partial pmoA and amoA protein sequences from
public databases and translated from PCR-based clone libraries from late successional
deciduous forest sub-plots. The tree is based on 164 amino acid positions using Phylip
Protein Maximum Likelihood as implemented in ARB (26). Bolded clades were
recovered from Rothamsted Research soils. Beneath each bolded clade is a listing of the
number of species within that clade found in which Rothamsted Research treatment:
Knott Wood (Wood), Broadbalk Wilderness (Wilderness) and Broadbalk Wheat (Wheat).
The scale bar represents 10 PAM units. ............................................................ 95

Figure 3.8. Similarity of methanotroph communities in paired sites featuring long-term
fertilized and non-fertilized soils. (A) The Sarenson index was calculated for each
pairwise comparison of methanotroph species using two-dimensional non-metric
multidimensional scaling (21). (B) The same data as in (A), but clustered using neighbor-

joining with MEGA (35) and displayed as a dendogram. Labels with a 7:? are sites that
have been fertilized long-term. Additional site information is provided in Table 3. 1 . .96

Figure A.1. Expected and actual relative abundance of the artificial pmoA and amoA
community Mix 2. ................................................................................. 115

Figure B]. Ammonia and nitrate concentrations in fertilized and control Late DF sub-
plots at KBS LTER. Error bars represent the standard error. ............................... 120

Figure C.1. Similarity of methanotroph communities in soils from around the globe. The
dendogram is based on Serenson index calculations for each pairwise comparison of the
methanotroph communities using pmoA species (defined as pmoA sequences having 94%
average nucleotide sequence similarity), and then clustered using neighbor-j oining with
MEGA (48). The scale bars represent at 0.1 change in the Serenson index. Additional site
information is provided in Table C.l ............................................................ 126

xiii

Chapter 1

Uncertainties in the Global Methane Budget and
Biological Influences on the Concentration of Atmospheric Methane

Introduction

To better understand the soil methane sink and its capacity for atmospheric
methane consumption, the studies in this thesis attempt to link variation in methane flux
associated with different land uses to changes in the richness and composition of
communities of methane consuming bacteria (methanotrophs). We then investigate the
effect of long-term fertilization, one of the factors associated with the land use changes,
to determine if it is associated with methanotroph community changes. As an
introduction to these studies, this chapter reviews the role of the soil methane sink in the
global methane budget, the methane consuming bacteria that are responsible for the
methane sink, the microbial pathway and enzyme that facilitates methane oxidation, the
effects of changing land use on the methane sink, and an overview of the following
chapters.

The Concentration of Atmospheric Methane

Methane (CH4) is a potent greenhouse gas whose atmospheric concentration, as

of 2005, is 1774 ppb (1). Methane contributes approximately 15% to the atrnosphere’s
total radiative forcing (1), and is 21-25 times more efficient at trapping heat (infrared
radiation) than carbon dioxide due to a longer lifetime in the atmosphere and greater
radiative efficiency (2). The present concentration of atmospheric methane represents its
highest concentration in at least the past 650,000 years (1). Since pre-industrial times

(ca. 1900), the concentration has increased 250% largely due to human agricultural

practices and fossil fuel use (1). Following almost a decade with little change in
atmospheric methane concentrations, as of 2007, there are renewed increases measured at
all worldwide monitoring stations (3).

The balance between methane emissions and the strength of the largest methane
sink, photochemical oxidation by hydroxyl radicals in the troposphere, primarily
determines the concentration of atmospheric methane (4). Therefore, it is thought that
the current increase in methane is either due to increases in wetland emissions in Siberia,
or a weakening of the hydroxyl radical sink (3). However, ambiguity in our knowledge
of multiple facets of the global methane budget has resulted in our inability to definitively
identify the cause of either the current increase or the decade of little change in the
concentration of atmospheric methane (3). Many other methane sources apart from
wetlands, the largest source of emissions, could be changing the magnitude of their flux,
and the strength of other methane sinks could also be changing.

Sources of Atmospheric Methane

There are numerous environments that are methane sources (yield a net
production of methane to the atmosphere). Many anaerobic environments produce
methane due to the activity of methanogenic archaea, and their activity accounts for more
than 70% of methane emissions. These environments include wetlands, landfills, oceans,
domestic and wild ruminant animals, poorly drained soil, and termites (5). Other sources
of methane include methane hydrates, wildfires, fossil fuel mining and use, and biomass
burning (5). Despite the identification of so many methane sources, the existence and

magnitude of emissions from additional methane sources is currently debated.

For example, plants producing methane under aerobic conditions was reported by
Keppler et al. (6), and originally estimated to be contributing between ~10-40% of annual
methane emissions. Critiques of their methods and data fi'om ice cores revised their
estimates downwards to 0-10% of total annual emissions (7, 8). Attempts to replicate the
initial observation have been inconsistent, with some studies confirming the initial
observation of methane emission from plants (9-12), while others have been unable to
corroborate those findings (13-15). Some studies have reported that the observed
methane is created through UV degradation of plant pectin (9, 11, 12), and if shown to be
the causal mechanism it would also explain the inability of some studies to replicate the
initial findings as UV light was not included in experimental conditions (13, 14).
However, the level of UV light exposure that causes methane emissions due to pectin
degradation far exceeded the amount of natural UV light exposure (11), and normal UV
light conditions would not yield large methane emissions from plants (13). Therefore, the
contribution of plants to methane emissions remains controversial.

Another possible contributor to methane emissions that is not typically accounted
for in current methane budgets is methane emissions originating from geological sources.
The methane is produced from microbial and therrnogenic processes in Earth’s crust and
subsequently released into the atmosphere through faults, seepage, and other
mechanisms. Current measurements have considerable error associated with them, but
they are estimated to possibly contribute 6% of total annual emissions (5). Thus, two
potentially significant sources of methane emissions remain controversial, and the
magnitude of their flux and their contributions to levels of atmospheric methane remain

unconstrained.

Methane Sinks

The number of methane sinks (environments that yield a net uptake of methane
from the atmosphere) is far fewer than the number of methane sources, but there is
similar uncertainty in the quantification of their impact on the concentration of
atmospheric methane. Approximately 90% of the methane consumed by methane sinks
is due to the photochemical oxidation by hydroxyl radicals in the troposphere. The other
methane sinks, stratospheric loss and uptake by well-drained soils, split the remaining
10% of methane sink consumption (5).

The hydroxyl radicals in the troposphere are very short lived, and producing
accurate measurements of the strength of the sink and its variability remains challenging
despite recent methodological improvements (4). Consequently, accurate modeling of the
troposphere sink and the ability to attribute changes in the concentration of atmospheric
methane as an effect of changes to the troposphere sink remains difficult (3). In addition,
factors causing variations to the hydroxyl radical sink are unknown. Due to the strength
of the troposphere sink, modifications to it will have large implications for the
concentration of atmospheric methane. Like other methane sources and sinks, future
investigations will have to clarify the role of the troposphere sink in detemrining the
concentration of atmospheric methane.

Compared to the troposphere sink, the atmospheric methane consumed by well-
drained soils is small: 30 z 15 Tg of methane, roughly 3-9% of the total amount of
methane removed from the atmosphere (16). However, well-drained soils are the only
biological sink of methane (5). The uptake of methane in well-drained soils is due to the

direct consumption by aerobic methane-oxidizing bacteria (methanotrophs) in soils

(Reviewed by (17)). Methanotrophs are ubiquitous, and will consume a significant
portion of the methane produced in most of the environments that are methane sources
(i.e. (18)), but only in well-drained soils does their activity yield an environment with a
net uptake of methane (see comment above). In a reaction mediated by their methane
monooxygenase enzyme (MMO), methanotrophs consume methane via an initial

oxidation with oxygen to produce methanol with the following stoichiometry:

CH4 + 02 + NADH + H+ —> CH3-OH + H20 + NAD+
After several chemical conversions (Figure 1.1), the carbon from methane is either

incorporated into biomass or respired as C02.

Aerobic Methanotrophs

Aerobic methanotrophs are phylogenetically distinct bacteria, and are found
within the Proteobacteria and Verrucomicrobia phyla. Verrucomicrobia methanotrophs
have only recently been described, having been isolated from extremely acidic aquatic
environments (19-21). Their optimum methanotrophic activities are typically in the
range of pH 2.0-2.5 (21), are capable of oxidation down to pH 0.8-1 (20), and do not
contain the intracellular membranes typical of Proteobacteria methanotrophs (19). '
Although one strain has had its genome sequenced (22), our knowledge of these
methanotrophs remains limited. Thus far, Verrucomicrobia methanotrophs have only
been discovered from aquatic environments, and Verrucomicrobia methanotrophs have
not been found in soils.

Isolates of Proteobacteria methanotrophs have been well studied in culture, but
the importance of cultured isolates to the consumption of atmospheric methane and the

soil methane sink is limited (see discussion below). Proteobacteria methanotrophs are

found in the gamma-Proteobacteria and alpha-Proteobacteria subdivisions. Type I
methanotr0phs are gamma-Proteobacteria, and include the genera Methylomonas,
Methylobacter, Methylococcus, and Methylomicrobium. Type II methanotrophs are
alpha-Proteobacteria, and include the genera Methylosinus and Methylocystis. In
addition to their subgroup classifications, Type I and Type II Proteobacterz'a
methanotrophs can be distinguished on the basis of the following traits: dominant
phospholipid fatty acids, serine verses RuMP carbon assimilation pathways, cellular
morphology, ability to fix nitrogen, G+C content of DNA, and the ability to form resting
stages (reviewed in (23)). Notably, some authors define Methylococcus capsulatus
(Bath) and similar strains as Type X methanotrophs due to the presence of both serine
and RuMP carbon assimilation pathways, and possession of traits similar to Type II
methanotrophs.
Other Methane Oxidizing Microbial Communities

Anaerobic methane oxidation has also been observed, and like aerobic methane-
oxidation it can attenuate a significant portion of the methane emitted from methane
sources (24). Details of metabolism are still being unraveled, but the process is
attributed to methanotrophic archaea (ANME) who are most likely performing reverse
methanogenesis within a microbial consortia (reviewed by (24)). ANME have typically
been found in close association with sulfate-reducing bacteria (25), but other bacteria also
appear to be involved in the microbial consortia (26). Denitrifying bacteria
anaerobically oxidizing methane without ANME has also been reported (27), further
highlighting the possibility that other electron acceptors could be involved (24). ANME

have thus far only been found in environments that are net methane sources and due to

their requirement for anoxia, it is unlikely that they contribute to the consumption
atmospheric methane in well-drained soils.

Ammmonia-oxidizing bacteria have also been thought to potentially contribute to
the strength of the soil methane sink as MMO is evolutionarily related to ammonium
monooxygenase (AMO) (28), and both enzymes are capable of oxidizing a wide range of
substrates that includes the other enzyme’s substrate (reviewed in (23)). However, while
AMO is capable of oxidizing methane, its Km for methane is much higher than MMO’s
Km, such that about a thousand ammonia-oxidizers are required to achieve the same rate
of CH4 oxidation as a single methanotroph (29). No evidence has been found for
ammonia oxidizers playing a significant role in methane oxidation under low or high
meflrane levels in either microcosms or the environment (29-31). Therefore, while it is
possible for ammonia-oxidizers oxidize to methane, their in situ contribution to
atmospheric methane consumption is, at best, minimal. As a result, we can assume that
the strength of the soil methane sink is mostly, if not entirely, due to the activity of
aerobic methanotrophs.

Methane Monooxygenase

Methane monooxygenase (MMO) is the key enzyme in the oxidation of methane,
and is found in both soluble and particulate forms. In the methanotrophs that produce
both forms, soluble MMO (sMMO) is expressed under low copper conditions, but there
is sufficient copper available in nearly all environments that it is rarely expressed in the
environment. Conversely, the particulate MMO (pMMO) is found in all known
methanotrophs (32) except two Methylocella strains (33). Due to the near ubiquity of the

particulate MMO in methanotrophs, the gene encoding the A subunit of pMMO, pmoA, is

typically used for non-culture based assessments of the methanotmph community in the
environment (17). Methanotrophs can be also be identified through their pmoA gene due
to its phylogeny being congruous with their 16S ribosomal genes (17). 168 ribosomal
genes can also be used for the identification of environmental methanotrophs, but many
of the primer pairs will capture methylotrophs in addition to methanotrophs (reviewed by
(17)). In addition, use of pmoA is advantageous for methanotroph detection instead of
ribosomal genes because unlike 16S sequences, the inference of methanotrophy is not
constrained by our phylogenetic knowledge of cultured methanotrophs. For instance, if
a novel pmoA is found it can be assumed to be a novel methanotroph, but if a novel 16S
gene is discovered, the assumption of methanotrophy is uncertain.

pMMO is made up of three subunits encoded by the genes, pmoC, pmoA and

. . . 70
pmoB, Wthh are found consecutively m an operon under the control of a a promoter

(34). There are typically two nearly identical copies of the pmoCAB operon in the
genome (34, 35). pMMO is an integral membrane protein, and as a result it has proven
difficult to study as purified protein preparations are unstable with low specific activities
(32, 36). Therefore, our knowledge of pMMO’s biochemistry remains limited and
controversial. The pMMO active site remains unknown despite solved protein structures
(37, 38), and there are numerous hypotheses as per the location and nature of the active
site. The active site could potentially lie in any of the three subunits that make up pMMO
(subunits A, B and C), and data supports the active site metal center as either, or some
combination of: diiron (36), mononuclear (3 7), dinuclear (3 7, 38) or trinuclear copper

metal centers (39, 40). Further hampering our understanding of pMMO is the continued

inability to correlate the reaction of substrates or products with any of the hypothesized
active sites (reviewed by (41)).
Methanotrophs Responsible for Consumption of Atmospheric Methane

Regardless of its biochemistry or the role it plays in catalyzing the oxidation of
methane, the pmoA gene remains an effective method of characterizing and identifying
the methanotroph community. Surveys of pmoA in well-drained soils have rarely found
the well-studied cultivable Proteobacteria methanotrophs detailed above (reviewed in
(17)). These observations were not completely unexpected, as the inability of cultured
methanotrophs to grow on atmospheric methane had led to the hypothesis that uncultured

“high affinity” methanotrophs are responsible for the strength of the soil methane sink

(42). Most of the cultured methanotrophs have a half-saturation constant (Km) that is too

high to allow grth on atmospheric methane. Only three Methylocystis strains (sp.
DWT, sp. LR] and sp. SC2) have been reported to approach the required atmospheric
methane oxidation kinetics, and to be capable of prolonged survival at atmospheric
methane concentrations (43-45). For at least Methylocystis sp. SC2, and in all probability
the other strains, their ability to use low methane concentrations is due to a pMMO
isozyme (43). Methylocystis sp. SC2 has two pmoCAB operons, each encoding a
different pMMO, and Baani and Liesack (43) demonstrate that survival at low methane
concentrations (10-100 ppm) is due to one isozyme, while growth at high methane
concentrations (>600 ppm) is due to the other isozyme.

However, while the strains have been reported to survive for up to three months
under low methane concentrations, no growth has been observed when they have been

incubated under atmospheric methane concentrations (1774 ppb; 1.74 ppm) (29, 43 -46).

It is possible that Methylocystis strains are contributing to the consumption of
atmospheric methane in well-drained soils by surviving via atmospheric methane
oxidation for maintenance energy in between exposures to higher concentrations of
methane that allow for its growth. West and Schmidt (47) found support for such a
possibility, as they were able to stimulate methanogenesis after exposing a well-drained
arctic tundra soil to anaerobic conditions. Upon the soil’s subsequent return to aerobic
conditions atmospheric methane was consumed at a faster rate; indicating that the
exposure of methanotrophs to greater methane concentrations is advantageous to the
methanotrophic community and the soil methane sink. Thus, so long as the Methylocystis
strains are occasionally exposed to greater levels of methane produced by methanogens
stimulated via anaerobic conditions from anoxic soil microsites or occasional soil
flooding, they will be able to persist for a prolonged time on atmospheric methane and
contribute to the strength of the soil methane sink.

While these frndings indicate that at least some cultured methanotrophs are
contributing to the oxidation of atmospheric methane they are not likely to be the
methanotrophs responsible for the majority of atmospheric methane oxidation. Instead,
methanotrophs that have not yet been cultured are hypothesized to likely be responsible.
Multiple studies looking at soils from throughout the world using culture independent
molecular surveys have found as yet uncultured novel numerically dominant and
phylogenetically distinct pmoA in well-drained soils (29, 46, 48-61). These pmoA
sequences form distinct phylogenetic clusters apart from the cultured gamma- and alpha-
Proteobacteria and Verrucomicrobia methanotrophs (Figure 1.2), and a selection of their

names are: JRl, JR2, JR3, MR1, Cluster I, Cluster II, Cluster III, Cluster V, Upland Soil

10

Cluster-y (USC-y or WBSF-H), Upland Soil Cluster-or (U SC-or or RA14), and RA21 (44,
46, 48-50, 55, 57).

As the sequences from these clusters are numerically dominant and commonly
found in well-drained soils throughout the world, it is plausible to conclude that they are
the methanotrophs responsible for the majority of biological atmospheric methane
oxidation. However, it has not yet been demonstrated that any of these clusters can grow
on atmospheric methane. Kolb et al. (29) found in situ pmoA expression for USC-or in a
German forest soils, but could not detect Cluster I in situ pmoA expression. Cluster I is
the only one of the clusters that has cultured representatives, but the strains have only

been minimally characterized (55). None of the exact methane oxidizing capabilities

(Km, Vmax, etc.) of any of the above clusters is known, and therefore a determination of

the magnitude of their contribution to atmospheric methane oxidation awaits further
supporting data.
Linking Methane Consumption to the Methanotroph Community

Although a full understanding of atmospheric methane oxidation will require the
culturing of many uncultured methanotrophs, our ability to measure methanotroph
diversity through the retrieval of the pmoA gene allows for the linking of methanotroph
diversity and community structure to rates of methane consumption. An example of a
model system where the link between the methanotroph community and methane
consumption can be tested is in well-drained soils of the same soil type that differ in their
land use.

The soil methane sink is significantly affected by land use with conversion of

soils to agricultural use, based on worldwide measurements, leading to an approximately

11

70% reduction in net methane consumption, and an approximately 100 year recovery to
the methane consumption rates of the native land use (reviewed by (62)). Land use
conversion to agricultural land-use introduces many environmental changes, and studies
examining total methane flux have found that fertilization, tillage, pesticide and herbicide
application, changes in water filled pores space, dry bulk density, and pH can attenuate
methane consumption in agricultural soils (reviewed in (63, 64)). In particular,
fertilization is an acute disturbance to methanotrophs. Ammonia is a known inhibitor of
MMO (63, 65, 66), and long-term application of fertilizer negatively impacts the
methanotroph community (30, 54).

However, paired sites that differ in land use share many environmental
characteristics like climate and soil type that better facilitates our ability to identify
factors that are impacting the methanotroph community and rates of methane
consumption due to conversion to agricultural soils. An example of one such paired site,
and the one used to investigate the methanotroph community in this thesis is the Kellogg
Biological Station Long Term Ecological Research Site (KBS LTER;
http://lter.kbs.msu.edu).

KBS LTER features a range of different agricultural treatments as well as an
entire successional gradient (early, mid and late successional soils). There are at least
three replicates of each land use, and for many of the treatments in situ measures of
methane consumption have been regularly made between March-December since 1992.
Like other paired sites, the magnitude of the soil methane sink at KBS LTER is affected
by land use. Rates of methane oxidation increase along a successional gradient:

conventional row-crop agriculture soils consume the least amount of methane, mid-

12

successional fields have intermediate rates of methane consumption, and late

successional forests have the highest rates. On a yearly average, the late successional

forest soils consume 9.17 g CH4-C ha'1 day-1, roughly 6 times more methane than the

conventional row-crop agricultural soils (1.62 g CI-I4-C ha'1 day-1) (67).

The robust gas measures, the large rate differences between the soils of the
successional gradient, and replicated plots make KBS LTER an ideal choice for exploring
the link between methanotroph diversity and community structure to rates of methane
consumption. In addition, KBS LTER features sub-plots in the late successional forest
that allow for the determination of how one factor associated with row-crop agricultural
management, long-term fertilization, affects methane consumption and the methanotroph
community.

Previous studies have not tried to correlate methanotroph richness to rates of
methane consumption as methane consumption rates have been well examined in soils of
varying land use (reviewed by (62)), but studies of the methanotroph community in
comparable sites have typically only taken a limited number of rates measurements (46,
53, 56, 59, 60). The results from these studies have been inconclusive with no clear
relationship observed between methanotroph richness and methane consumption (46, 53,
68). Methanotroph community changes have been observed across other successional
gradients, with patterns of methanotroph diversity changing along the successional
gradients of reforested comfields (46) and reclaimed pasture lands (68). KBS LTER
features more land use types in its successional gradient then either of those study sites

and coupled with robust gas measurements can better resolve patterns between the

13

methanotroph community, rates of methane consumption, and land use then any previous
study.
Summary

In order to fully understand the causes of the increase in atmospheric methane,
there are many areas of the global methane budget that require greater understanding.
One such area is the soil methane sink, the only biological methane sink, where the
activity of aerobic methanotrophs results in the net consumption of atmospheric methane.
The methanotrophs that are probably responsible for the consumption of atmospheric
methane have not yet been cultured, and a range of non-cultured methanotrophs have
been revealed in culture-independent molecular surveys. Land use changes, particularly
the conversion of native lands to row-crop agriculture, result in a dramatic drop in the
rate of methane consumption, but changes in land use have not been linked to changes in
methanotroph community diversity and structure. Paired sites of differing land use and
changed rates of methane consumption, like those featured at KBS LTER, represent ideal
model systems for linking methanotroph diversity to rates of methane consumption.
Thesis Overview

The main goals of this thesis is to (1) determine how the methanotroph
community changes along with rates of methane consumption, (2) to begin to investigate
factors that may be causing the changes, and (3) to determine if our findings at KBS
LTER are applicable to other sites. By doing so, insights can be gained into how lands
might be managed to enhance the methanotroph community, and in turn, the capacity of
soil methane sink.

Therefore, the main overall questions of this thesis were:

14

(a) Do soils of various land uses at KBS LTER, with different rates of methane
consumption, harbor different methanotroph communities?

(b) Can the differences in the methanotroph community be directly related to
changes in the rate of methane consumption?

(c) At KBS LTER is a decline in methanotroph richness associated with long-term
fertilization alone?

(d) Are there typical changes to methanotroph communities in response to long-
term fertilization?

To answer questions (a) and (b), chapter 2 presents the findings of assessments of
the methanotroph community from the successional gradient at KBS LTER We find
that the conversion of native lands to row-crop agriculture causes the loss of
methanotroph richness, and a correlated decrease in the rate of methane consumption.
We also find that land uses harbor different microbial communities that appear to change
in parallel to differences in the plant community.

To determine the importance of methanotroph diversity to rates we find that the
recovery of both methanotroph richness and methane consumption is linear, concurrent
and will take approximately 75 years following abandonment fi'om agriculture. The
linear trajectory and lack of a step-wise increase in methane consumption rates following
any increase in methanotroph diversity suggests that every methanotroph taxon, and not
just a few taxons, are contributing to the rate of methane consumption at the KBS LTER
Further supporting the finding of complementarity within the methanotroph community is
the comparison of correlations of methane consumption with methanotroph richness,

carbon dioxide efflux with total bacterial richness, and of both gas fluxes with moisture

15

and temperature. We find that carbon dioxide efflux, a process that is the result of a
highly redundant microbial community does not correlate with total bacterial richness,
and that more of the carbon dioxide flux can be explained by moisture and temperature.

To answer questions (c) and (d), in chapter 3 I present the findings from
assessments of the methanotroph community from KBS LTER late successional forest
fertilized sub-plots, from agricultural and forest soils from the Rothamsted Research site,
and from previously published studies of methanotroph communities. We find that long
term fertilization alone is not causing a decrease in methanotroph richness nor in rates of
methane consumption at KBS LTER This result is consistent with our findings in
Chapter 2 that linked methanotroph richness and rates of methane consumption, and led
to the expectation that the response of richness and rates of methane consumption to
long-term fertilization would be the same. No consistent decline in methanotroph
richness is observed in other long term fertilized soils, and methanotroph communities
display a distinct biogeography with communities clustering together based on
geographic location. There is no pattern of typical methanotroph community changes in
accordance with long-term fertilization.

Based on these findings we conclude that the each methanotroph taxon at KBS
LTER provides a similar contribution to rate of methane consumption, and with every
additional methanotroph in the soil there is an increase in the rates of consumption. The
causes behind the change in richness remain unclear, as they are not caused by long-term
fertilization alone, and other agricultural land-use associated changes are implicated as
helping to cause the decrease in methanotroph richness in the Ag HT treatment at KBS

LTER. The results also indicate that the findings at KBS LTER may not be directly

16

applicable to other soils possibly due to each site’s unique methanotroph community and
soil properties. Further experiments that would clarify the effects of land use on the
methanotroph community are discussed in Chapter 4.

Three appendices are also provided. One details the PCR bias associated with the

reaction conditions used in this study, and how the finding led to exclusion of abundance

measures in the methanotroph community analyses. The second appendix presents NH4+

and N03' nutrient data fi'om a fertilized late successional forest sub-plot at KBS LTER

before and after fertilization. The sub-plot’s methanotroph community was not assessed,
but the findings reinforce the expectation that the methanotrophs in the fertilized sub-
plots are exposed to ammonia following fertilization, and highlight the active nitrification
in KBS LTER late successional forest soils. The third appendix confirms the finding of
methanotroph biogeography in long-term fertilized soils by assessing the clustering of

methanotroph community composition in a variety of well-drained soils.

17

.6552: 95 0:28.: .6 :osmExo 05 $558.: 533 .ommgwxxoocofi
052:2: 2: m_ 0::an m8— of. detects 2856:. ..o .3358 0:832: 05 .8 ocmfiozom “BE—:86 < .2 8:9“.

 

 

 

mmanE 33:8 3.: wouaeomoofi ON: NC
w 0% + m TEIIV
O ommgwobigew EOOOEA ommaewouwznow OEUEA omwswegsow :0 mo emacowbnoo:oz EU

Bane—om efiEeEagom 35522 89502

18

a—proteobacteria

    
 
 
 
 
 
 

Cluster V

 

   

[Upland Soil Cluster 0493‘]

 

y-proteobacteria amoA

y—proteobacteria Type I methanotrophs
Cluster Ill

Upland Soil Cluster 7

V

 

 

 

=Venuoomicrobia pmoA
cggnnmr n «cum ll '
mm
Unknown
Classification
/
fi-proteobacten’a amoA
Crenothrix polyspora) y—proteobacteria

\
Cluster ll Presumably a—proteobaclen'a

Cmnarachaeota amoA

Figure 1.2. Phylogenetic tree of selected partial pmoA and amoA protein sequences from
public databases. The tree is based on 164 amino acid positions using Phylip Protein
Maximum Likelihood as implemented in ARB (Ludwig et al. 2004). Boxed labels are
indicative of pmoA clades that have been found in soils, have not been cultured, and are
thought to likely play a substantial role in the oxidation of atmospheric methane. The
scale bar represents 10 PAM units.

19

References

1.

10.

IPCC (2007) Climate Change 2007: Synthesis Report. Contribution of Working
Groups I, II and III to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. eds Team CW, Pachauri RK, & Reisinger AGeneva,
Switzerland).

Forster P, et al. (2007) Changes in Atmospheric Constituents and in Radiative
Forcing. Climate Change 200 7: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the Intergovernmental Panel
on Climate Change, eds Solomon S, Qin D, Manning M, Chen Z, Marquis M,
Averyt K, Tignor M, & Miller H (Cambridge University Press, Cambridge), pp
129-234.

Rigby M, et al. (2008) Renewed Growth of Atmospheric Methane. Geophysical
Research Letters 35:doi:10.1029/2008GL036037.

Bousquet P, et al. (2006) Contribution of anthropogenic and natural sources to
atmospheric methane variability. Nature 443(7110):439-443.

Denman KL, et al. (2007) Couplings Between Changes in the Climate System and
Biogeochemistry. Climate Change 200 7: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, eds Solomon S, Qin D, Manning
M, Chen Z, Marquis M, Averyt K, Tignor M, & Miller H (Cambridge University
Press, Cambridge), pp 499-587.

Keppler F, Hamilton If G, Brass M, & Rockmann T (2006) Methane emissions
fiom terrestrial plants under aerobic conditions. Nature 439(7073): 1 87-191.

Kirschbaum MUF, et al. (2006) A comment on the quantitative significance of
aerobic methane release by plants. Functional Plant Biology 33(6):521-530.

Ferretti DF, et al. (2007) Stable isotopes provide revised global limits of aerobic
methane emissions from plants. Atmospheric Chemistry and Physics 7:237-241.

Keppler F, et al. (2008) Methoxyl groups of plant pectin as a precursor of
atmospheric methane: evidence fi'om deuterium labelling studies. New Phytologist
178(4):808-814.

Wang Z-P, Han X-G, Wang GG, Song Y, & Gulledge J (2008) Aerobic Methane

Emission from Plants in the Inner Mongolia Steppe. Environmental Science &
Technology 42(1):62-68.

20

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Vigano I, et al. (2008) Effect of UV radiation and temperature on the emission of
methane from plant biomass and structural components. Biogeosciences 5(3):937-
947.

McLeod AR, et al. (2008) Ultraviolet radiation drives methane emissions fiom
terrestrial plant pectins. New Phytologist 180(1):124-132.

Kirschbaum MUF & Walcroft A (2008) No detectable aerobic methane efflux
from plant material, nor from adsorption/desorption processes. Biogeosciences
5(6):1551-1558.

Beerling DJ, Gardiner T, Leggett G, McLeod A, & Quick WP (2008) Missing
methane emissions from leaves of terrestrial plants. Global Change Biology
14(8):1821-1826.

Dueck TA, et al. (2007) No evidence for substantial aerobic methane emission by
terrestrial plants: a 13C-labelling approach. New Phytologist l75(1):29-35.

Wuebbles DJ & Hayhoe K (2002) Atmospheric methane and global change.
Earth-Science Reviews 57(3 -4):l77-210.

McDonald IR, Bodrossy L, Chen Y, & Murrell JC (2008) Molecular Ecology
Techniques for the Study of Aerobic Methanotrophs. Appl. Environ. Microbiol.
74(5):1305-1315.

Chanton JP, Powelson DK, & Green RB (2009) Methane oxidation in landfill
cover soils, is a 10% default value reasonable? Journal of Environmental Quality
38(2):654-663.

Islam T, Jensen S, Reigstad LJ, Larsen 0, & Birkeland NK (2008) Methane
oxidation at 55 degrees C and pH 2 by a thermoacidophilic bacterium belonging
to the Verrucomicrobia phylum. Proceedings of the National Academy of
Sciences of the United States of America 105(1):300-304.

Pol A, et al. (2007) Methanotrophy below pH 1 by a new Verrucomicrobia
species. Nature 450(7171):874-878.

Dunfield PF, et al. (2007) Methane oxidation by an extremely acidophilic
bacterium of the phylum Verrucomicrobia. Nature 450(7171):879-882.

Hou SB, et al. (2008) Complete genome sequence of the extremely acidophilic
methanotroph isolate V4, Methylacidiphilum infemorum, a representative of the
bacterial phylum Verrucomicrobia. Biology Direct 3.

Hanson RS & Hanson TE (1996) Methanotrophic bacteria. Microbiological
Reviews 60(2):439-+.

21

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Caldwell SL, et al. (2008) Anaerobic oxidation of methane: mechanisms,
bioenergetics, and the ecology of associated microorganisms. Environmental
Science & Technology 42(18):6791-6799.

Moran JJ, et al. (2008) Methyl sulfides as intermediates in the anaerobic oxidation
of methane. Environmental Microbiology 10(1): 162-173.

Pemthaler A, et al. (2008) Diverse syntrophic partnerships from-deep-sea

methane vents revealed by direct cell capture and metagenorrrics. Proceedings of
the National Academy of Sciences of the United States of America 105(19):7052-
7057.

Ettwig KF, et al. (2008) Denitrifying bacteria anaerobically oxidize methane in
the absence of Archaea. Environmental Microbiology 10(1 1):3164-3173.

Holmes AJ, Costello A, Lidstrom ME, & Murrell JC (1995) Evidence that
participate methane monooxygenase and ammonia monooxygenase may be
evolutionarily related. FEMS Microbiology Letters l32(3):203-208.

Kolb S, Knief C, Dunfield PF, & Conrad R (2005) Abundance and activity of
uncultured methanotrophic bacteria involved in the consumption of atmospheric

methane in two forest soils. Environmental Microbiology 7(8):1150-1161.

Gulledge J, Hrywna Y, Cavanaugh C, & Steudler PA (2004) Effects of long-term
nitrogen fertilization on the uptake kinetics of atmospheric methane in temperate
forest soils. FEMS Microbiology Ecology 49(3):389-400.

Carini SA, Orcutt BN, & Joye SB (2003) Interactions between methane oxidation
and nitrification in coastal sediments. Geomicrobiology Journal 20(4):355-3 74.

Kitrnitto A, Myronova N, Basu P, & Dalton H (2005) Characterization and
structural analysis of an active particulate methane monooxygenase trimer from
Methylococcus capsulatus (Bath). Biochemistry 44(33):]0954-10965.

Dunfield PF, Khmelenina VN, Suzina NE, Trotsenko YA, & Dedysh SN (2003)
Methylocella silvestris sp nov., a novel methanotroph isolated from an acidic
forest carnbisol. International Journal of Systematic and Evolutionary
Microbiology 53: 123 1-1239.

Gilbert B, et al. (2000) Molecular analysis of the pmo (particulate methane
monooxygenase) operons from two type II methanotrophs. Applied and
Environmental Microbiology 66(3):966-975.

Ward N, et al. (2004) Genomic insights into methanotrophy: The complete

genome sequence of Methylococcus capsulatus (Bath). Plos Biology 2(10):1616-
1 628.

22

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

Martinho M, et al. (2007) Mosssbauer studies of the membrane-associated
methane monooxygenase from Methylococcus capsulatus bath: Evidence for a

diiron center. Journal of the American Chemical Society 129(5 1): 15783-+.

Lieberman RL & Rosenzweig AC (2005) Crystal structure of a membrane-bound
metalloenzyme that catalyses the biological oxidation of methane. Nature
434(7030):177-182.

Hakemian AS, et al. (2008) The metal centers of particulate methane
monooxygenase from Methylosinus trichosporium OB3b. Biochemistry
47(26):6793-6801 .

Chan SI, et al. (2007) Redox potentiometry studies of particulate methane
monooxygenase: Support for a trinuclear c0pper cluster active site. Angewandte
Chemie-International Edition 46(12): 1 992-1 994.

Chan SI & Yu SSF (2008) Controlled oxidation of hydrocarbons by the
membrane-bound methane monooxygenase: The case for a tricopper cluster.
Accounts of Chemical Research 41(8):969-979.

Rosenzweig AC (2008) The metal centres of particulate methane mono-
oxygenase. Biochemical Society Transactions 36: 1 134-1 137.

Bender M & Conrad R (1992) Kinetics of CH4 Oxidation in Oxic Soils Exposed
to Ambient Air or High CH4 Mixing Ratios. FEMS Microbiology Ecology
101(4):261-270.

Baani M & Liesack W (2008) Two isozymes of particulate methane
monooxygenase with different methane oxidation kinetics are found in
Methylocystis sp strain SCZ. Proceedings of the National Academy of Sciences of
the United States of America 105(29): 10203-10208.

Knief C & Dunfield PF (2005) Response and adaptation of different
methanotrophic bacteria to low methane mixing ratios. Environmental
Microbiology 7(9): 1307-13 17.

Dunfield PF & Conrad R (2000) Starvation alters the apparent half-saturation
constant for methane in the type H methanotroph Methylcystis strain LRl.
Applied and Environmental Microbiology 66(9):4136-4138.

Knief C, et al. (2005) Diversity of methanotrophic bacteria in tropical upland soils
under different land uses. Applied and Environmental Microbiology 71 (7):3826-
3 83 1 .

West AE & Schmidt SK (2002) Endogenous methanogenesis stimulates oxidation
of atmospheric CH4 in alpine tundra soil. Microbial Ecology 43(4):408-415.

23

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

Holmes AJ, et al. (1999) Characterization of methanotrophic bacterial populations
in soils showing atmospheric methane uptake. Applied and Environmental
Microbiology 65(8):3312-33 18.

Henckel T, Jackel U, Schnell S, & Conrad R (2000) Molecular analyses of novel
methanotrophic communities in forest soil that oxidize atmospheric methane.
Applied and Environmental Microbiology 66(5): 1 801-1808.

Horz H-P, Rich V, Avrahami S, & Bohannan BJM (2005) Methane-Oxidizing
Bacteria in a California Upland Grassland Soil: Diversity and Response to
Simulated Global Change. Appl. Environ. Microbiol. 71(5):2642-2652.

Jaatinen K, Knief C, Dunfield PF, Yrjala K, & Fritze H (2004) Methanotrophic
bacteria in boreal forest soil after fire. FEMS Microbiology Ecology 50(3): 195-
202.

King GM & Nanba K (2008) Distribution of Atmospheric Methane Oxidation and
Methanotrophic Communities on Hawaiian Volcanic Deposits and Soils.
Microbes and Environments 23(4):326-330.

Knief C, Lipski A, & Dunfield PF (2003) Diversity and activity of
methanotrophic bacteria in different upland soils. Applied and Environmental
Microbiology 69(1 l):6703-6714.

Lau E, Ahmad A, Steudler PA, & Cavanaugh CM (2007) Molecular
characterization of methanotrophic communities in forest soils. FEMS Microbiol
Ecol 60(3):490-500.

Pacheco-Oliver M, McDonald IR, Groleau D, Murrell JC, & Miguez CB (2002)
Detection of methanotrophs with highly divergent pmoA genes from Arctic soils.
FEMS Microbiology Letters 209(2):3 13-3 19.

Reay DS, Radajewski S, Murrell JC, McNamara N, & Nedwell DB (2001) Effects
of land-use on the activity and diversity of methane oxidizing bacteria in forest
soils. Soil Biology & Biochemistry 33(12-13): 1613-1623.

Ricke P, Kolb S, & Braker G (2005) Application of a newly developed ARB
software-integrated tool for in silico terminal restriction fragment length
polymorphism analysis reveals the dominance of a novel pmoA cluster in a forest
soil. Applied and Environmental Microbiology 71(3): 1671-1673.

Singh BK & Tate K (2007) Biochemical and molecular characterization of

methanotrophs in soil from a pristine New Zealand beech forest. FEMS
Microbiology Letters 275(1):89-97.

24

59.

60.

61.

62.

63.

65.

66.

67.

68.

Singh BK, et al. (2007) Effect of afforestation and reforestation of pastures on the
activity and population dynamics of methanotrophic bacteria. Applied and
Environmental Microbiology 73(16):5 153-5 161 .

Zhou XQ, et al. (2008) Effects of grazing by sheep on the structure of methane-
oxidizing bacterial community of steppe soil. Soil Biology & Biochemistry
40(1):258-261.

Knief C, Kolb S, Bodelier PLE, Lipski A, & Dunfield PF (2006) The active
methanotrophic community in hydromorphic soils changes in response to
changing methane concentration. Environmental Microbiology 8(2):321-333.

Smith KA, et al. (2000) Oxidation of atmospheric methane in Northern European
soils, comparison with other ecosystems, and uncertainties in the global terrestrial
sink. Global Change Biology 6(7):791-803.

Hutsch BW (2001) Methane oxidation in non-flooded soils as affected by crop
production - invited paper. European Journal of Agronomy 14( 4):23 7- 260.

King GM (1997) Responses of atmospheric methane consumption by soils to
global climate change. Global Change Biology 3(4):351-362.

Gulledge J & Schimel JP (1998) Low-Concentration Kinetics of Atmospheric
CH4 Oxidation in Soil and Mechanism of NH4+ Inhibition. Appl. Environ.
Microbiol. 64(1 1):4291-4298.

Suwanwaree P & Robertson GP (2005) Methane Oxidation in Forest,
Successional, and No-till Agricultural Ecosystems: Effects of Nitrogen and Soil
Disturbance. Soil Science of America Journal 69: 1722-1 779.

Robertson GP, Paul EA, & Harwood RR (2000) Greenhouse Gases in Intensive
Agriculture: Contributions of Individual Gases to the Radiative Forcing of the
Atmosphere. Science 289: 1922-1925.

Zhou XQ, Wang YF, Huang XZ, Tian JQ, & Hao YB (2008) Effect of grazing
intensities on the activity and community structure of methane-oxidizing bacteria
of grassland soil in Inner Mongolia. Nutrient Cycling in Agroecosystems
80(2):l45-152.

25

Chapter 2

Agriculture’s Impact on Microbial Diversity and
the Flux of Greenhouse Gases

Abstract

Row-crop agriculture impacts both the production of carbon dioxide and the
consumption of methane by microbial communities in upland soils - Earth’s largest
biological sink for atmospheric methane. To determine if there are relationships between
the rates of these ecosystem level processes and the diversity of microbes that catalyze
them, we measured the in situ fluxes of methane and carbon dioxide, and conducted
molecular surveys of methane—oxidizing bacteria (methanotrophs) and total bacterial
diversity across a range of land uses. Rates of methane consumption and methanotroph
diversity were positively correlated, as conversion of lands to row-crop agriculture led to
a 7-fold reduction from maximal values found in native deciduous forests. In fields
abandoned from agriculture the diversity of methanotrophs and the consumption of
methane increased monotonically, suggesting that managing lands to conserve or restore
methanotroph diversity could help mitigate increasing atmospheric concentrations of this
potent greenhouse gas. In addition, methanotroph diversity was more capable at
explaining methane consumption than was moisture and temperature. Conversely, total
bacterial diversity did not correlate with changes in carbon dioxide emission, but did
correlate to moisture and temperature. Taken together, these results are consistent with
the prediction that ecosystem processes are more likely to be influenced by microbial

diversity when microbial communities have limited species richness.

26

Introduction

Carbon dioxide (C02) and methane (CH4) are responsible for approximately 80%

of the positive radiative forcing of the atmosphere from long-lived greenhouse gases
(approximately 62% and 18%, respectively) (1). The atmospheric concentrations of both
gases now exceed any of their respective levels over the past 65 0,000 years, and continue
to rise (2, 3). Fluxes of both gases are affected by land-use changes — especially
deforestation and row-crop agriculture (4, 5). Conversion of native upland soils (e.g.
forest, grassland) to agricultural management reduces their capacity for methane
consumption by an average of 71%, and recovery of methane consumption is estimated to
be ca. 100 years following the cessation of agriculture (reviewed in (5)). Deforestation
and agricultural development increases the efflux of carbon dioxide fiom soil (4, 6), and
is estimated to have contributed approximately 25% of the increase in radiative forcing
from carbon dioxide since 1850 (1).

Microbes directly control methane consumption, and portions of soil carbon
dioxide efflux (7, 8). Aerobic methane-oxidizing bacteria (methanotrophs) consume
methane in well-drained soils, and the consequential net uptake of methane constitutes
Earth’s largest known biological sink for atmospheric methane (9, 10). Soil respiration is
a large biological source of atmospheric carbon dioxide, and is mainly a reflection of
microbial respiration (reviewed in (11)) plus respiration from roots. The relative
contribution of root and heterotrophic microbial respiration has proven difficult to
disentangle (12). Nevertheless, changes in soil respiration have been associated with

variations in the microbial community (8, 13).

27

How row-crop agriculture changes the methanotroph and the heterotrophic
microbial community and their associated rates of methane consumption and carbon
dioxide efflux remains unclear. The Kellogg Biological Station Long Term Ecological
Research Site (KBS LTER), with replicated plots of the same soil series (14) that just
differ in land use (Table 2.1), represents an ideal system for determining how row-crop
agriculture impacts microbial communities and the greenhouse gas fluxes they control.
Past studies have typically investigated either gas flux or the microbial community at
analogous sites, but rarely has the function mediated by the microbes and the microbial
community been investigated simultaneously in agricultural soils or across a land-use
gradient. We correlated in situ fluxes of carbon dioxide and methane to assessments of
microbial diversity across a successional gradient featuring 5 different land uses at KBS
LTER (Table 2.1).

Methane consumption rates have been well examined in soils of varying land use
(reviewed by (5)), but studies of the methanotroph community in comparable sites have
either not measured rates of methane consumption (15), taken a maximum of two rate
measurements (16-19), not featured replicated sites (16, 18, 19), and have only examined
a limited number of successional stages (15-19). No clear relationship has been
observed between methanotroph richness and methane consumption in any study (18-20),
and, to our knowledge, ours is the first study that has attempted to correlate methanotroph
richness to rates of methane consumption. Methanotroph community changes have been
observed across other land-use gradients, with patterns of methanotroph diversity
changing along a gradient of grazing intensities (20), and along a successional gradient of

reforested comfields (18). KBS LTER features more land use types in its successional

28

gradient then any previous study, and coupled with robust gas measurements can better
resolve patterns between the methanotroph community, rates of methane consumption, '
and land use.

Similarly, few studies have attempted to determine the effects of row-crop
agriculture on both soil respiration and the heterotrophic microbial community. Studies
have documented land-use associated changes to the microbial community due to the
conversion of forest to pasture (20, 21), but rates of soil respiration were not measured.
Others have linked changes in soil respiration to changes in the microbial community (8,
13, 22), but agricultural soils were not included in any of those studies.

We hypothesize that microbial diversity is more likely to be important to a
specialized metabolic process like methane oxidation than it is to be a factor in a
metabolically redundant process like carbon dioxide production (23-25). Specialized
metabolic processes like methane consumption are typically the result of relatively
taxonomically narrow microbes, and a limited number of species contribute to the
ecosystem process, resulting in a lack of fimctional redundancy. On the other hand, soil
respiration is a process that is the result of a multitude of taxonomically diverse microbes
and enzymatic processes, and the number of microbes found to be contributing to carbon
dioxide production is likely to be high. Therefore, we expect that methanotroph richness
will correlate to rates of methane consumption, but total bacterial richness will not
correlate to carbon dioxide efflux.

In addition, by measuring methanotroph diversity and methane consumption in
fields that had been abandoned from agriculture for various lengths of time, we should to

be able to infer if the postulated positive relationship between net methane consumption

29

and methanotroph diversity is due complementarity or selection. The complementarity
and selection hypotheses are actively debated to explain positive relationships between
the magnitude of an ecosystem process and species richness (26, 27). The
complementarity hypothesis gives importance to every species in determining the
magnitude of the ecosystem process, and postulates that increased ecosystem function
results from the combined activity of species in complementary niches. The selection or
sampling effect hypothesis offers an alternative explanation: the positive relationship
between diversity and function is driven by one or a few dominant species particularly
proficient at performing the process under study. Discerning between the two hypothesis
is important because if the postulated positive relationship between net methane
consumption and methanotroph diversity is observed, it will be useful to identifying
whether all methanotroph diversity, or only select species need to be conserved or
restored in order to possibly enhance the methanotroph community, and in turn, the

capacity of the soil methane sink.

Materials and Methods
Site Description

This study was conducted at the Kellogg Biological Station Long Term
Ecological Research Site (KBS LTER; http://lter.kbs.msu.edu) located in Hickory
Comers, Michigan. Soils of five treatments were examined in this study: conventional
agricultural management of historically tilled land (Ag HT), early successional plant
communities on fields that had been abandoned from agriculture in 1989 (Early HT),

mid-successional plant communities on either historically tilled land (Mid HT) or never

30

tilled grassland (Mid NT), and a never tilled late successional deciduous forest (Late DF).
Additional site descriptions can be found in Table 2.1. All of the KBS LTER soils are
located within 3 km of the main experimental site, are all of the Kalamazoo/Oshtemo soil
series, and are well-drained Typic Hapludalfs (fine or course loamy, mixed).
Rate Measurements

In situ rates of methane consumption were measured using closed-cover flux
chambers (28). Rate measurements were typically made twice monthly between March
and December at 3 or 4 replicate plots between 1992 and 2007 for Ag HT, Early HT and
Late DF, and between 1992 and 1997 for Mid NT and Mid HT. Mid HT rate
measurements were also made in 2002. Individual rate measurements and detailed
methods are available at http://lter.kbs.msu.edu. To determine rate differences based on
treatments, analysis of variance (AN OVA) was performed using PROC MIXED (SAS
Inc, 2002).
Soil Sampling

Methanotroph diversity was assessed in 5 soil cores (2.5 x 10 cm) collected fi'om
3 replicates along a gradient of 5 land uses (Table 2.1). These 75 samples, collected from
a total of 15 experimental fields on 13 June 2006, as well as samples used for additional
molecular surveys collected on 8 December 2004, 6 June 2005, and 13 June 2007 were
transported to the laboratory on ice where they were mixed thoroughly and flash frozen in

a dry ice/ethanol bath, then stored at -80°C until processing.

Total bacterial diversity was assessed in 5 soil cores (5 x 10 cm) from two Ag HT
and two Late DF replicates in December 2006. Samples were pooled, transported to the

laboratory on ice, sieved and stored at -80°C.

31

DNA Extraction and pmoA PCR Reactions

For the assessment of total bacterial diversity, DNA was extracted according to
Zhou et al. (29), followed by a cesium-chloride gradient purification (30). For the
assessment of methanotroph diversity, DNA was extracted from soil samples with the Mo
Bio PowerSoilTM DNA Isolation Kit, following the manufacturer’s protocol, except that
mechanical cell lysis was performed by bead beating for 45 seconds.

All soil samples were screened for genes coding for the A subunit of particulate
methane monooxygenase (pmoA) via PCR amplification with the primer sets A189 (5’-
GGNGACTGGGACTTCTGG-B’) -A682 (5’-GAASGCNGAGAAGAASGC-3’) (31),
and A189-mb66l (5’-CCGGMGCAACGTCYTTACC-3’) (32) in order to encompass all
known pmoA genes (3 3). No amplification was observed from A189-mb661.

Amplification reactions contained either 25, 45 or 90 ng of undiluted DNA, 1.25
ul 1% BSA, 10 mM dNTPs, 0.2 uM of each primer, 2.5 ul 10x PCR buffer (200mM
Tris-HCl (pH 8.4) and 500 mM KCl), 0.5 pl 50 mM MgC12 and 1.5 U of T aq DNA
polymerase (Invitrogen, Carlsbad, CA) in a total volume of 25 ul. Reaction conditions

for the 62°C-56°C clone libraries and for tRF LP (see below) were 95°C for 5 minutes,
15 cycles of a ‘touchdown PCR’ of 95°C for 1 minute, 62°C for 1 minute (-O.4°C each
cycle to 56°C), and 72°C for 1 minute, 15 cycles using 56°C as the annealing

temperature, and a final 10 minute extension at 7 2°C. To ensure that no methanotroph

diversity was missed, additional libraries were constructed under the same conditions, but

with varying annealing temperatures: 60°C to 51°C (-0.6°C each cycle to 51°C), 48°C,
and 51°C (Table 2.2).

Clone Libraries

32

Cloning was performed with the TOPO-TA cloning kit (Invitrogen, Carlsbad,
CA) using either vector pCR 4 or pCR 2.1 as per the manufacturer’s protocol.
Transformants were screened via PCR reactions with the primers F2 (5 ’-
CAGTCACGACGTTGTAAAACGACGGC-3’) and R4 (5’-
CAGGAAACAGCTATGACCATG-3’) (34). One-1.25 ul of the PCR products were

purified via incubation with 0.25 ul of ExoSAP-IT (U sb, Cleveland, 0H) for 30 minutes

at 37°C. Sequencing was completed at the Research Technology Support Facility at

Michigan State University (RTSF).

Sequences identified by BLASTX (3 5) as pmoA or the A subunit of ammonia
monoxygenase (amoA), which A189-A682 also amplify, were imported into Arb (36). In
Arb, sequences were translated and aligned using Clustal W. Nucleic acid sequences
were aligned according to the protein sequence. Sequences from clone libraries were
determined to be the same species if they were 2 94% identical (3 7) as determined by
DOTUR (average neighbor grouping) (3 8).

Soil cores collected in December 2004 and June 2005 were used to construct

62°C-56°C clone libraries from 5 cores from the l”t replicates of Ag HT and Late DF.

For June 2005 samples, libraries represent sequences pooled from separate DNA
extractions, PCR reactions and cloning reactions. These libraries also include clones
whose identities were inferred from identical banding patterns in restriction fragment
length polymorphism analyses using the Alwl enzyme (N eb, Ipswich, MA). The same

June 2005 soil samples were also used to construct libraries from 48°C, 51 °C, and 60°C-

51°C annealing temperatures. For these annealing temperatures, separate DNA

33

extractions were pooled for 4 replicate PCR amplifications, which were subsequently
pooled into 1 cloning reaction.

Libraries fiom June 2006 samples were constructed from duplicate 60°C-51°C

PCR reactions from each of 5 soil cores from the lSt and 3rd Late DF and 1st and 2lfld Ag
HT replicates which were pooled according to replicate. Prior to cloning these PCR
products, as well as those from June 2007, were digested with the restriction enzyme
PflFI (N eb, Ipswich, MA), and gel extracted with the PrepEase kit (U sb, Cleveland, OH)
to reduce the incidence of cloning amoA and non-specific PCR products.

Additional 62°C -56°C clone libraries were constructed from a collection of the

June 2006 soil cores from various treatments where large unidentified peaks were
observed in the tRF LP analysis (see below). The libraries were constructed in order to
match more tRF LP peaks with known sequences, and therefore were not sampled until
pmoA rarefaction curves were asymptotic.

Sequences representing each of the pmoA species found in this study are
deposited in GenBank under accession numbers FJ529724 - FJ529808, and GQ219582-
GQ2195 83.

Terminal Restriction Fragment Length Polymorphism (tRFLP) Analysis

For tRF LP (39) each of the 75 soil cores from June 2006 was amplified for pmoA
with A682 labeled with fluorophore 6-carboxyfluorescein (6-F am), and template addition
was always 45 ng. To obtain enough DNA for tRF LP at least 4 replicate PCR reactions
were pooled. PCR products were purified using a MinElute column (Qiagen, Valencia,
CA). In a 50 u] reaction, 300ng of purified product was digested with 1.5 U Taul

(F ermentas, Glen Burnie, MD) by incubating at 55°C for 1 hr 30 min. To inactivate the

34

enzyme the DNA was precipitated as follows: The sample was diluted to 500 it],
followed by the addition of 50 ul 3M sodium acetate, 1 ul or 2.5 pl 10 mg/ml glycogen,
and 500 pl isopropanol, and holding on ice for at least 5 minutes. The DNA was then
pelleted by centrifugation at 16,000 x g for either 5 (1 ul glycogen) or 10 minutes (2.5 ul
glycogen). The supernatant was decantated, and the pellet washed with 500p] of 80%
ethanol followed by centrifugation for 2 minutes at 16,000 x g, and removal of the
supernatant. After a 30 second centrifugation additional ethanol was removed, and the
DNA was air dried for 5-10 minutes before resuspension in 20 ul of water. In an 18-22
ul reaction, either 140 ng (lul glycogen) or 160 ng (2.5ul glycogen) of DNA was
digested with 2.5 U SspI (Neb, Ipswich, MA) in a 1hr 30min incubation at 37°C. After
heat inactivation at 65°C for 20 minutes 6 U of BstUI (Neb, Ipswich, MA) was added,

incubated at 60°C for 1hr 30min, and inactivated by adding 0.8 ul of 0.5 mM EDTA.

Capillary electrophoresis of the tRF LP reactions was then performed with a 5 fu cutoff at
RTSF.

To ensure that the tRFLP profiles could be compared, the distribution of the total
fluorescence was compared with PROC UNIVARIATE (SAS Inc, 2002), and any
outliers were excluded. Individual peaks were distinguished from the background signal
and binned using TRFLP-Stats (40). In TRFLP-Stats default settings were used except
for the standard deviation cutoff, which was increased to 4.5. The resulting cutoff of
approximately 25 fu excluded any profile whose highest peak area was not greater than
350, and ensured that peak areas below 300 were not identified as true peaks in greater
than 90% of the samples. PCR products from clones representing 14pmoA OTUs, 1

amoA, and 21 non-specific (neither pmoA or amoA) were run as tRF LP at least once.

35

These clone controls allowed the identification of specific tRF LP peaks as either pmoA
OTUs, amoA, or non-specific PCR products, and allowed for the exclusion of amoA and
non-specific PCR product bins. Those bins that were identified as either were excluded
from the analysis. To obtain a better representation of an entire replicate, and to control
for soil heterogeneity, bins from the same replicate were summed. For every replicate,
tRFLPs from at least 3 soil cores were obtained and summed. In total, tRF LP profiles
were obtained from 68 of the 75 soil cores.

At least two negative controls (no DNA PCR reactions) were included on each
plate of tRF LP reactions, and any profile whose highest peak area was not greater than
300, the highest peak area in these controls, was excluded from the analysis. To assess
the reproducibility of the method, twenty samples had technical replicates whose results
were then pooled after analysis in TRFLP-Stats.

Each tRF LP bin then served as an Operational Taxonomic Unit (OTU), and the
OTUs were used in a linear regression (PROC REG, SAS Inc, 2002) against the rate of
methane consumption at KBS-LTER in summer (J une-August), against the rate of
methane consumption between March-August, and to calculate B-diversity with PAST
(41) or Estimate S (http://viceroy.eeb.uconn.edu/EstimateS).

168 Tag Sequencing

16S tagged sequencing and analysis was performed on a 454 Life Science’s
pyrosequencer as described by Sogin et al. (42). Richness estimates were then regressed
against mean carbon dioxide measures taken between 2005-2007 using simple linear
regression (PROC REG, SAS Inc, 2002).

Multiple regression

36

To determine the ability of soil moisture and maximum air temperature to predict
rates of methane consumption and carbon dioxide emission, we used multiple regression
(PROC MIXED, SAS Inc, 2002) with data taken between 2005-2007 from Late DF and
Ag HT. Moisture and soil temperature measurements were taken on the same day that
gas flux was measured. Individual measurements are available at

http://lter.kbs.msu.edu/datasets.

Results
Rate measurements

Maximal rates of methane consumption and carbon dioxide emission were
observed during the summer months for all treatments (Figure 2.1). Carbon dioxide
emissions peaked May through August, and significant treatment effects were found
(Figure 2.2a). The conventional row-crop agricultural soil (Ag HT) had the lowest
average rate of soil respiration, and the early (Early HT) and mid-successional sites (Mid
NT and Mid HT) had maximal rates. On average, the early and mid-successional sites
emit approximately 79% more, and Late DF emits 18% more, carbon dioxide than Ag HT
per day.

Rates of methane consumption also changed according to treatment (Figure 2.2b)
with the greatest consumption in Late DF followed by the mid-successional soils, Early
HT, and the lowest rates in Ag HT (Figures 2.1 and 2.2b). The overall difference in rates

between the highest rate in Late DF and the lowest rate in Ag HT is approximately 7-fold
(9.98 g CH4-C ha'1 clay'1 and 1.29 g CH4-C ha'1 day-1, respectively), resulting in 7 times

more methane being consumed per day in the late successional deciduous forest as

37

compared to the conventional row-crop agricultural soils. The difference between the
Late DF and Ag HT treatments grew 2-fold with the inclusion of rates from 2000-2007
(28). The highest rates of methane consumption were observed June through September
(Figure 2.1), the 7-fold difference in rates remained the same during those months, and
coincided with when we obtained the majority of the soil samples used to assess the
methanotroph community.
Methanotroph Diversity and Community Composition
pmoA Clone libraries

Clone libraries of the pmoA gene, encoding the A-subunit of the particulate
methane monooxygenase, the first enzyme in the pathway of methane oxidation and the
defining enzyme of aerobic methanotrophs (43), were used to assess methanotroph
diversity. Libraries from Late DF and Ag HT were constructed from various annealing
temperatures, soil samples (Methods, Table 2.2), and until rarefaction curves were
asymptotic (Figure 2.3) to ensure that all methanotroph diversity had been captured. The

libraries constructed fi'om the various annealing temperatures revealed that the 60°C-
51°C annealing temperature yielded the most methanotroph species in an individual

library (Table 2.2). Due to PCR bias, measures of pmoA abundance were excluded from
comparisons of the methanotroph community (Appendix A).

Although the A-subunit from some ammonia monooxygenase genes (amoA)
amplified with the pmoA primers (31), amoA sequences were distinguished based on
diagnostic amino acids (44, 45), and their clustering in phylogenetic trees (Figure 2.4,
Figure 2.5). Digestion of the PCR products with the restriction enzyme PflFI (N eb,

Ipswich, MA) was found to greatly reduce the incidence of cloning amoA, and to have no

38

change in the overall methanotroph richness recovered from individual libraries (Table
2.2).

Phylogenetic analysis of pmoA genes revealed that methanotrophs in KBS LTER
soils cluster within seven clades (Figure 2.4). Six pmoA clades - Cluster 1, Cluster II,
KBSl , JRl, MR1, and Upland Soil Cluster a - were found in Late DF, compared with
just 2 clades, Cluster I and Cluster II, in Ag HT. Cluster RA21 was only recovered from
Mid NT (Figures 2.4). Grouping the pmoA sequences from the KBS LTER clone
libraries at a species level (94% average nucleotide similarity) revealed an even more
dramatic difference in the methanotroph richness between the Ag HT and Late DF land-
use treatments: 24 methanotroph species in Late DF, and only 7 species were present in
Ag HT (Table 2.2, Figure 2.5). Of the 7 methanotroph species found in Ag HT, 2 were
unique to the treatment while the other 5 species were also found in Late DF . In both
Late DF and Ag HT, more methanotroph species were found in Cluster I then in any
other cluster, with 12 Late DF and 6 Ag HT species. Next was Cluster II with 7 Late DF
and 1 Ag HT species. The other clusters had no more than 2 methanotroph species.
Cluster KBSl was represented by 1 Late DF sequence fiom the libraries in this study, and
additional sequences from the KBSl clade are reported from a clone library constructed

from a subplot of the lSt Late-DF replicate that was sampled on 13 June 2007 with 60°C-
51°C as the annealing temperature (Chapter 3) to further confirm the presence of this

unique cluster (Figure 2.5).
Terminal Restriction Fragment Length Polymorphism (tRFLP) Analysis
To survey methanotroph diversity across the entire gradient of KBS LTER land

uses (Table 2.1), and in three replicates of each treatment, terminal Restriction Fragment

39

Length Polymorphism (tRF LP) of the pmoA gene was used with soil sampled during the
peak period of methane consumption in 2006. The tRF LP assay distinguished 11
Operational Taxonomic Units (OTUs) that were confirmed as pmoA gene fiagments
through sequence analysis, and the requirement for comigration with terminal restriction
fragments from cloned controls. Similar to the clone libraries, richness differences
between treatments were observed. The Late DF sites averaged 7 OTUs compared to 2
OTUs in Ag HT, and intermediate numbers of pmoA OTUs were found in the
successional sites (Figure 2.6).

In addition, following release from row-crop agriculture (Ag HT), the
composition of the methanotroph communities in soils abandoned from agriculture (Early
HT and Mid HT) became more like the native sites (Late DF and Mid NT) (Figures 2.8
and 2.9). These methanotroph communities are statistically different from one another
according to a one-way analysis of similarities (p < 0.005, ANOSIM), with
methanotrophs in the early successional soil beginning to diverge from the row-crop
agricultural plots. Divergence continued in the Mid HT soils that have been abandoned
from agricultural for ca. 50 yrs., such that these communities began to overlap in
composition with communities of methanotrophs in native fields (Figures 2.8 and 2.9).
Correlation of methanotroph diversity and methane consumption

Simple linear regression revealed a strong positive correlation between summer

(June-August) rates of methane consumption and pmoA OTUs (r2=0.62, p<0.001) (Figure

2.6). If pmoA OTUs are regressed against rates of methane consumption from March-

December a nearly identical relationship is found (r2=0.64, p<0.001). The same strong

positive correlation is also exhibited when, in addition to the pmoA OTUs, tRF LP OTUs

40

that could not be identified (Methods) are included and correlated with summer (r2=0.43,

p=0.008), and March-December methane consumption rates (r2=0.48, p=0.004).

We tested whether the positive relationship between methane consumption and
methanotroph diversity at the KBS LTER was more likely to be explained by either the
complementarity or selection hypotheses by determining the trajectory of recovery of
richness and consumption in successional soils (Early HT and Mid HT) after intensive

row-crop agricultural management (Figure 2.7). There was a linear trajectory for the

recovery of both methane consumption (r2=0.69, p < 0.001) and methanotroph richness

(r2=0.99, p = 0.020) over time. Extrapolation of the trajectories reveals that if methane

consumption and methanotroph diversity continue at the same rates since cessation of
agriculture, both would return to the current level of their equilibrium community, the
late successional deciduous forest, in approximately 75 years.
Total Bacterial Diversity

Estimated average total bacterial richness (Chao I) was determined fi'om 454 168
tag sequencing fi'om December 2006 soil samples, and found to be 11,105 in Ag HT, and
7,762 in Late DF . We assumed that every bacterium would contribute to carbon dioxide
emissions, and preformed a linear regression with total bacterial richness and average

March-December carbon dioxide efflux from 2005-2007. No relationship was found

between the two measures (r2=0.22, p=0.522) (Table 2.3). Despite the limited power in

the linear regression due to the limited number of samples (n=4), using the same number
of samples from comparable methanotroph richness measurements yields a dramatically

different result (Table 2.3). Using pmoA clone libraries from Ag HT and Late DF June

41

2006 soil samples (Table 2.2), and average methane consumption measurements from the
same 2005-2007 dates, we found a significant positive linear relationship between

average March-December methane consumption measures and methanotroph diversity
(r2 =0.96, p = 0.013).

Multiple Regression

Multiple regression with the Ag HT and Late DF gas fluxes from 2005-2007
against soil moisture and maximum air temperature from the same dates and replicates
discerned the ability of these general microbial metabolic regulators to explain the
observed fluxes. Moisture and temperature were found to be able to explain portions of

both fluxes, but were 10 times more effective in explaining the efflux of carbon dioxide

(r2 =0.37, p < 0.001) then the rates of methane consumption (r2 =0.03, p = 0.005) (Table

2.3).

Discussion

Microbial Diversity and the Flux of Greenhouse Gases

Correlating ecosystem process rates to their respective microbial diversity
measures across the KBS LTER land-use gradient conformed to our hypothesis: methane
consumption rose as methanotroph diversity increased (Figure 2.6, Table 2.3), and
increases in carbon dioxide emission were not correlated to increases in bacterial richness
(Table 2.3). The lowest flux of both greenhouse gases were found in Ag HT as both
carbon dioxide and methane consumption fluxes were diminished by row-crop
agriculture (Figures 2.1 and 2.2). The difference in the correlations was instead due to

the effect of land use on microbial diversity: Row-crop agriculture was associated with a

42

dramatic decrease in methanotroph richness, but total bacterial richness was greater in Ag
HT then it was in Late DF (Figure 2.6, Tables 2.2 and 2.3).

As postulated by our hypothesis, the ability of microbial diversity to explain the
rate of the gas flux is likely the result of the functional redundancy of the microbial
community mediating the ecosystem process. Thousands of bacterial species contribute
to the efflux of carbon dioxide fi'om soil — reducing the likelihood that diversity
influences the process rate. There are comparatively few methanotroph species, which
likely results in little or no functional redundancy, as methanotrophs probably occupy
non-overlapping niches and every methanotroph species contributes to the rate of
methane consumption.

Evidence for methanotrophs occupying separate niches at KBS LTER was gained
by testing to see if the positive relationship between net methane consumption and
methanotroph diversity could best be explained by either the complementarity or
selection hypotheses. Ideally this determination would be able to be made with measures
of species-specific methane oxidation rates and the reconstruction of defined
communities, but the absence of cultured methanotrophs representative of the clades
found in KBS LTER soils preclude such measures. As has been found in other soils, the
methanotroph community consuming atmospheric methane at KBS-LTER is composed
of uncultured methanotrophs. Our findings further indicate that these phylogenetic
clusters are those that are largely responsible for the consumption of atmospheric
methane (10, 19). No clones were found from the Type I or Type H methanotrophs that
are well represented in culture (Figures 2.4 and 2.5). Rather, pmoA clones belonging to 6

clades that have been identified previously in other culture independent investigations of

43

upland soils (10, 18, 19, 45), and 1 new clade, identified as KBS], were found in KBS
LTER soils. Of the 7 clades that were recovered, only Cluster I methanotrophs,
presumably OC-proteobacteria, have been reported to be cultured, but they remain poorly
characterized (46).

Despite the absence of cultured methanotrophs, we detemiined whether the
complementarity or selection hypotheses could best explain the relationship between
methane consumption and methanotroph richness by plotting the recovery of both
measures in soils that had been abandoned from agriculture for various lengths of time.
The trajectory of the recovery of both methane consumption and methanotroph richness
over time is linear, with no indication of abrupt step-wise increases in methane
consumption that would be expected to accompany the establishment of a particularly
productive methanotroph species in accordance with the selection hypothesis (Figure
2.7). Instead, the concurrent and incremental recovery of methane oxidation and
methanotroph diversity following the abandonment fi'om row-crop agriculture is
consistent with complementary roles of methanotroph species, and suggests that every
methanotroph OTU is important to rates of methane consumption.

Additional support for complementarity in the methanotroph community at KBS

LTER is found in the ability for general abiotic microbial metabolic regulators, moisture

and temperature, to explain 10 times more of the variation in C02 flux then it did for

methane consumption (r2=0.37 and 0.03, respectively; Table 2.3). Without an influence

of diversity on rates of soil respiration, the ability of moisture and temperature to exert
controls on the efflux of carbon dioxide was more apparent. In contrast, with diversity

explaining most of the rate of methane consumption, moisture and temperature could

44

only explain a minimal amount of the variation in the rate of methane consumption
(Table 2.3).

Other studies have similarly failed to find a correlation between total microbial
diversity (24, 47, 48) and soil respiration. These studies have also found that changes in
the microbial community were correlated to more metabolically specialized processes
like N20 production (24) and nitrification potential (24, 48). Studies that have not
directly measured total microbial diversity have also found support for the microbial
community not influencing carbon dioxide production, but influencing specialized
microbially mediated processes ((49), reviewed in (23, 25)). However, not all studies
have found that diversity always influences specialized physiological processes. Enwall
et al. (50) found rates of denitrification to be independent of microbial diversity, and
Wertz et al. (47) found similar results for denitrification and nitrification. There is also
evidence of microbial diversity being correlated to soil respiration (8, 13, 22).
Additionally, Bell et a1. (51) found bacterial respiration to be correlated with increasing
bacterial diversity in laboratory microcosms, but with a maximum of 72 bacterial species,
their ecosystem was likely not nearly as functionally redundant as the thousands of
bacterial species contributing to carbon dioxide production in soil.

Methanotroph richness changing with rates of methane consumption has also been
observed in other sites, but an overall pattern is unclear. In Mono Lake, the depth with
the highest rate of methane oxidation was found to have the greatest richness of
methanotrophs (52), and two pine forests had greater methanotroph richness and
methane consumption rates than paired pasture soils in New Zealand (16). However, that

same study found that a paired shrubland and pasture soil had the same methanotroph

4s

richness despite different rates of methane consumption, and other studies have been
inconclusive with no clear relationship observed between methanotroph richness and
methane consumption (18-20). Therefore, the relationships we have observed between
methane consumption and methanotroph diversity is consistent with some previous
findings, and the lack of an overall pattern may at least be partially due to methodological
differences. Most assessments of the methanotroph community are performed at a
broader phylogenetic level, with fewer replicates, and/or compared to less robust
measures of the rate of methane consumption. In addition, this is the first study to

correlate methanotroph richness to rates of methane consumption.

The Impact of Agriculture on Methanotroph Diversity

At KBS LTER, there is unambiguous evidence linking the diversity of aerobic
methanotrophs to the observed rates of methane consumption, and for agricultural
management to diminish methanotroph species richness. The 24 methanotroph species
that we recovered in Late DF (Figure 2.5, Table 2.2) is, to our knowledge, the most
methanotroph richness recovered from a single soil; other studies have reported between
1-13 methanotroph species (l6, 18, 19, 45). Conversion of soils to row crop agriculture
dramatically reduced that richness to a small subset of the methanotroph richness found
in Late DF.

In addition, the recovery of methanotroph diversity and methane consumption
after cessation of agriculture is projected to take approximately 75 years (Figure 2.7).
This slow recovery of methanotroph richness and methane consumption at the KBS

LTER is consistent with worldwide observations that suggest a recovery period of ca. 100

46

years for methane consumption following the cessation of agriculture (5). The lack of a
quicker recovery following the stopping of agricultural management practices, especially
those like fertilization that represent a disturbance to the methanotroph community,
indicate that many aspects of methanotroph niches are distressed due to row-crop
agriculture. It will be important to identify these pivotal variables and their applicability
to other sites if we are to manage lands to conserve or restore methanotroph diversity and
enhance the capacity of soil to serve as a sink for this potent greenhouse gas.

The methanotroph community changes across the successional gradient at KBS
LTER mirrors changes in the plant community, and further underscores the notion that
methanotroph niches are slowly re-established and colonized following row-crop
agriculture (Figures 2.8 and 2.9). There is a discemable pattern of recovery, and an
apparent succession of methanotrophs. Similar methanotroph community changes across
other successional gradients have also been observed, with patterns of methanotroph
diversity along successional gradients of reforested comfields (18) and reclaimed pasture
lands (20). In each of these studies, as well as our study, the methanotrophs changed
along with the plant community. King and Nanba (53) also found distinct methanotroph
communities in volcanic deposits with different plant communities. However, while
these observations suggest that plant diversity and/or community composition influences
methanotroph diversity, there is not an obvious causal connection between plant and

methanotroph diversity.

Conclusion

47

In conclusion, our results suggest that every methanotroph OTU is important to
methane consumption rates in KBS LTER soils, and both methane consumption and the
diversity of methanotrophs decline in response to row crop agriculture. The decline of
both methane consumption and methanotroph diversity in row-crOp agricultural soils, and
the long time required for recovery of methanotroph diversity suggests that multiple
aspects of the methanotrophs habitat are disrupted. It will be important to identify and
quantify the effect of these pivotal variables if we are to manage lands to conserve or
restore methanotroph diversity, and enhance the capacity of soil to serve as a sink for this
potent greenhouse gas. There is no relationship between soil respiration and bacterial
richness, and the contrasting result from methane consumption and methanotroph
richness is consistent with the prediction that microbial diversity is more likely to be
important to a specialized metabolic process whose a microbial commrmity is likely to be

of limited richness and consequently functionally redundant.

Acknowledgements
All methane rate measurements and were taken by G. Philip Robertson and members of
his laboratory. Soil sampling and DNA extraction for the assessment of total bacterial

diversity was performed by Dion Anotonopoulos.

48

.3828 SE8. 8.88 8 85 £828: 083. 8284—95 88 28 8:8 88280888823— 05 .8 08 £8 =<

3:82 820388 28 82.38% 288088 .8288 82:0 .8288 :898m-Eoo-~8:3
85:; E8688 2 28 .32 2 232288 89c noc¢u§n<

8288 32:58 2 2.8 28:: :89 .88: .83 .8 8.88: :8: ~88“—

ooiéma— .8253 22888 9.322%.“ Eat 8.88822

8:: 88 88: 82...: :8 "own 88» co. :8 88:9: 28
.88 soon 88: 822 88232 8:5 2: .8 83. .888 88:28“. 286888 83

am; :_ 8

m

:8.:mE.m€_.8~<\H&E 8 28$ up :8 288—8 82m 28222.

H: w< 88::otw< 2:288:00
H: bumm— _m=2m88:m Esm—
._.Z 22 82880 28588522
.5.— 22 .88"— 88888522
"5 33 ~88m 80:298.
88% 88

N. _ 48:8 25 E .8232ng 8:m MmEH—fimx .—.N 03:.

49

22 £238— .n0ZV 0850 862me Sam 05 .23 8888.0 083 0888 :8 085 89¢ 80:88 MD“:

50

8 v: 8:88,..- 8 0.3 33.5 058
5 8m 8838:8053
z «8 .8888 “E 838
v 8 3-8 38.6. 2:: N ,8 8
m 8 8.8 $58 2:: _ ,8 8
m 8 8 8m 2:: _ 8 0<
N 8 8-8 88 as: _ ,8 m...
_ mm 8-8 38 .2888 _ 8 0<
w 8 :8 88 2.3 m 8 33
o. a. 3-8 88 2.3 _ 8 33
E 8 :8 8m 2:; _ 8 93
8 3 8 88 as: _ ,8 23
a o: 8 8N 2:; _ 8 33
: S 8-8 88 2:: _ 8 23
v a 8-8 88 .3888 _ 8 23
858.888 . _ 5 8% all; 482 84:88 a .58.
$3888 85138.38 88888.8 88%

82.8.8: 08.0 «BER-8 gm .N.N 03:.

Table 2.3. Correlations between the flux of greenhouse gases, species richness and
environmental conditions.

g P Value
C02 Production]:
vs. bacterial richnessz 0.22 0.522
vs. temp and moisture4 0.37 <0.001
CH4 Consumption]:
vs. methanotroph richness3 0.96 0.018
vs. temp and moisture4 0.03 0.005

1 Rate measurements were typically made bimonthly between March and November of
2005 through 2007.

2 Linear regression (n=4) with total bacterial richness estimated with Chao I and based on
220,996 168 tag sequences detemiined from samples collected in December 2006. Rate
measures are arithmetic averages fi'om the same plots fiom which richness was
determined: The second and fourth replicates of Ag HT, and the first and third replicates
of Late DF.

3 Linear regression (n=4) with methanotroph richness determined at a 94% pmoA
sequence identity from clone libraries from samples collected in June 2006. Rate
measures are arithmetic averages from the same plots from which richness was
determined: The first and second replicates of Ag HT, and the first and third replicates of
Late DF.

4 Multiple regression (n=280) with maximum air temperature and soil moisture for the
day rates were measured. Rate measures are from four replicates of Ag HT, and from
three replicates of Late DF.

51

 

 

RR —7 T T i‘i:
p -2- i\§ i\ \i7!<g:§/§ i

7

i.’- \i\;7 #:H /
ii: ,/./
E ;/ ,

 

 

 

T T I l I I l l l l
Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Monlh

Figure 2.1. Average monthly carbon dioxide production and methane consumption based
on current land management and historical land use at the KBS-LTER: Agricultural
management of historically tilled land (Ag HT; V ), early successional plant communities
on fields that had been abandoned from agriculture in 1989 (Early HT; I ), mid-
successional plant communities on either historically tilled land (Mid HT; ) or never
tilled land (Mid NT; 0 ), or a late successional deciduous forest (Late DF; 0 ).

52

HO"

 

 

 

 

 

 

Ag HT Early HT Mid'HT Mid NT Late OF

B

 

Net Methane Consumption (g Cl-l4C ha" day")
0')
l
-l O

 

 

 

 

 

Ag HT Early HT Mid'HT Mid NT Late DF
Figure 2.2. The effect of different land uses on average carbon dioxide emission (a) and
net methane consumption (b) at KBS LTER. Difi‘erent letters represent significant
differences (p<0.05) between treatments. Rate measures are the same as those in Fig. 2.1.
Error bars represent standard errors. Land use treatments are: agricultural management of
historically tilled land (Ag HT), early successional plant communities on fields that had
been abandoned from agriculture in 1989 (Early HT), mid-successional plant
communities on either historically tilled land (Mid HT) or never tilled land (Mid NT), or
a late successional deciduous forest (Late DF).

53

50‘ LateDFRep1,Mnber'04,62°C—56°C
- -—-~LateDFRep1,June'05,62°C-56°C

 

 

45': —LateDFRep1,June'05,51°C
.3 4o— ——LateDFRep1,June'05,48°C
g _ —LateDFRep1,June'06,60°C—51°C

35_ — LateDF Rep 1, June '05,60°C-51°C
.: . — Late DF Rep 3, June '06, 60°051°C
g 30— —Agl-lTRep1,Deoerrber'04,62°C—56°C
8 - —Agl-l'l'Rep1,June'05,62°C—56°C
2 25- —Agl-lTRep1,June'05,48°C
u 4 —AgHTRep1,June'06, 60°C-51°C
g 20- —AgHTRep2, June'06, 60°C-51°C
E 15—
:E, 10—
Z .

5-

0 . l

 

' l ' I I l . I '

0102030405060708090100110
Number of Sequences

Figure 2.3. pmoA rarefaction curves from all of the KBS-LTER meA clone libraries
constructed from the Late DF and Ag HT treatments. Libraries were constructed from
DNA extracted from soil sampled in December 2004, June 2005, and June 2006 with
various annealing temperatures (For additional details see Table 2.2, and Methods). All
curves were constructed using data from neighbor joining matrixes from Arb (Ludwig et
al. 2004), with curves calculated by DOTUR (3 8). Methanotroph species are defined as
pmoA sequences having 94% average nucleotide sequence similarity. Error bars
representing 95% confidence intervals were omitted for the sake of clarity.

54

a—proteobacteria
' Upland Soil
Cluster a

y-proteobacteria
amoA

 

 

     
     
  
     

/ methanotrophs

y-proteobacteria Type l
methanotrophs

Upland Soil
Cluster 7

Verrucomicrobia
pmoA

fi Cluster ll @ Unknown
Classification

Unknown
Classification

fl-proteobacteria

o
_L
O

amoA Crenothn’x polyspora ) y-proteobacteria
Presumably
Crenalachaeota Clusmr I * a—proteobacteria
amoA

Figure 2.4. Phylogenetic tree of selected partial pmoA and amoA protein sequences from
public databases and translated from PCR-based clone libraries from KBS-LTER soils.
The tree is based on 164 amino acid positions using Phylip Protein Maximum Likelihood
as implemented in ARB (Ludwig et al. 2004). Boxed or circled labels are indicative of
pmoA clades recovered from KBS-LTER soils. Cluster I and Cluster II clades (boxed
and starred labels) were recovered from Ag HT and Late DF, while clades KBSl, IRl,
Upland Soil Cluster (1, and MR1 (boxed labels) were recovered from Late DF. Clade
RA21 (oval) was only recovered from Mid NT soil. The scale bar represents 10 PAM

55

Figure 2.5. Neighbor joining phylogenetic tree of the partial nucleic acid sequences of
pmoA and amoA from reference sequences, and KBS-LTER clone libraries from 'soil
sampled in December 2004, June 2005, June 2006, and June 2007. The tree is rooted
with the branching from the amoA sequences from Nitrosomonas eutropha and
Nitrosococcus mobilis. Nodes representing the 27 different methanotroph species
(defined as pmoA sequences having 94% average nucleotide sequence similarity)
identified at KBS-LTER are highlighted. The colors of the highlights reflect species’
membership in the seven pmoA clades depicted in Figure 5. The pmoA sequences
included in this tree are deposited in GenBank under accession numbers F J 529724 -
F1529808, and GQ219582-GQ219583.

56

Figure 2.5

 

Nitrosomonas europaea (AF037107)
NitrosococcuerriobZilisDe(AF03|7108)4
6D 0'0

I' L.,

9
Lt DF 2 De ' - '
a: B proteobacter'a amoA

 

Uncultured ancterium clone CL49 AJ699315
a. ' Late DF 51 8‘05 009 ( )

  

     
    
    
   
  
 

 

Uncultured bacterium DGGE band L58— b2 (AJ868244)

T6 6unJ 0612 SFS 30
T Late DF 43 June n05 H06

 

Lf uncultured bacterium clone p1LWPmoA- 24 (00067064)

Late DF 60- 51 June '06 B09
lncu tureo nactertum 0 one pmoAa4 (D0367742)

 

g T 8neJu H03- 2
Unc uult red bacterium DGGE band E33b- a (AJ579661)
LaC e DF 62- 56 Dec ‘04 E6
Late DF 60- 51 June '05 CO1
Ag HT 48Jun e'05 F01- 2

Li Ag HT 48 June '05 D07-2
W Uncultured bacterium clone 137233-79 (AJ564439) C|uster |
I ate DF 60- 51 June '06 801—1

L29 BE 895152152525:

Methanotrmh

1 Uncultured bacterium6 clene 4L7OPA1 2. 4 AF358041

2 ’Uncultured bacterium clone CL27 (AJ6 9310)
Late DF 51 June '05 D09

[:Late DF 51 June '05 (310-1

 

    
    
      
   
       
   

 

   
  

 

Late DF 60- 51 June '05 A06

Late DF 62- 56 June ‘051- 2- D11

Late DF 51 June ‘05 001-3

Late DF 6051 June '05 DOB

Late DF 51 June '05 E09-3

Late DF 48 June B-1 1

, Uncultured bacterium clone CL5 (AJ699316)
'7 ate DF 48 June '05 008-1

Late DF 48 June '05 803— 1

 

Late DF 51 June '05 CO4
!F n- HI

 

l
i
l
'l‘ Late DF 51 June '05 001 3

L re et a p -8 7175)
Methanotroph KS- 21 (AF547180)
,2 2,2, CrenothrI'x pol s are clone Cp moA- 12 (0025955693308)
L,” { CrenoLtJhr/X CUpo/yeydbpra clone CDp- éérrEioA- 2d(D(SQ2959 UCrenothr/x polyspora
bacterium E5FB- b (AJ579668 )

l J Meth gloca/dum sz zegedDiense U8930 \
Met y/ocaldum qrac/le (U89 O1)

 

57

 

 

Figure 2.5 Continued

   
     

0.10

 

 
 
 

Crenothrix polyspora
Type l methanotrophs

A AF5090 2)
amoAa’gXFS 9003 0 ))y—proteobacteria amoA

Type II methanotrophs

Upland Soil Cluster 0t

(AF148523)

  
  
  

JR1

Cluster ll

 
 

SK—c (AJ868278)

   

OF 62-56 June '06 1-2-42
DF 60—51 June '05 E11

   
  

MR1

RA2 1

 
 
 
  

bacterium 90_P96

V4 clone pmoA38 (EU2238 55) . _

bacterium SoImeoAi (EF 59108 6) Verrucomlcorbla pmoA
bacterium V4 clone pmoA2 (EU223862 )

 

58

 

 

..L
A

 

ha ’1 day '1)
on a re
I l I

Net Methane consumption (9 CH4-C
O)
I

 

 

 

 

T l j I I

r
5 6 7 8 9
Methanotroph richness (OTUs)

—""1

N—l
0)
.h

Figure 2.6. The correlation between methanotroph diversity and methane consumption at
KBS LTER. Relationship between summer methane consumption (June-August) and
methanotroph richness (both represent averages of 3 replicate plots) across landscapes at
the KBS LTER. A simple linear regression is presented (r2 =0.62, p <0.001) with
operational taxonomic units (OTUs) defined as peaks in the tRFLP analysis that have
been identified as a pmoA gene. Symbols are as follows: Agricultural management of
historically tilled land (Ag HT; V ), early successional plant communities on fields that
had been abandoned from agriculture in 1989 (Early HT; I ), mid-successional plant
communities on either historically tilled land (Mid HT; ) or never tilled land (Mid NT;
0 ), or a late successional deciduous forest (Late DF; 0 ).

59

 

4 A n

"' Ag Early Mid Forest ”

p Successional Successional F8
512—

'2 , A r7

a
a.

51"
t}:

Net Methane consumption (9 CH4-C
‘i’
n'a 0'») 4L
(sn 10) sseuuou udonoueuiew

I
.3

 

 

 

 

CL I I r f I l I I F
0 10 20 30 40 50 60 (69) (77)
Years since agricultural management

Figure 2.7. The recovery of methanotroph diversity and methane consumption at KBS
LTER following row—crop agriculture. Increase in methanotroph diversity (open
symbols) and methane consumption (closed symbols) as a function of time since
cesssation of agriculture. Measurements of the deciduous forest (Late DF) are positioned
based on projections from linear regression used to fit methanotroph diversity (y =0.07x
+2.05; r2=0.99, p = 0.020) or methane consumption (y=0.13x + 0.80; r2 =0.69, p < 0.001).
Error bars represent standard errors; symbols are as follows: Agricultural management of
historically tilled land (Ag HT; V ), early successional plant communities on fields that
had been abandoned from agriculture in 1989 (Early HT; I ), mid-successional plant
communities on either historically tilled land (Mid HT; ) or never tilled land (Mid NT;
0 ), or a late successional deciduous forest (Late DF; 0 ).

60

 

 

0.4-

0.3- ’

0.2-

o
N .
93 Native
3 0.1d l
E 9 Early Successiona-I
I
o 0 o A Agriculture
Mid Successional HTv
-0.1- A
. o
-o.2- A
‘0.3 I I

-0.3 -0.2 -O.1 O 0.1 0.2 0.3 0.4 0.5
Coordinate 1

Figure 2.8. Similarity of methanotroph communities among KBS-LTER treatments.

The Sarenson index was calculated for each pairwise comparison of methanotrophs using

OTUs for all confirmed pmoA s, and then plotted using two-dimensional non-metric

multidimensional scaling (Hammer et al. 2001). Symbols are as follows: Agricultural

management of historically tilled land (Ag HT; V ), early successional plant communities

on fields that had been abandoned fi'om agriculture in 1989 (Early HT; I ), mid-

successional plant communities on either historically tilled land (Mid HT; . ~. ) or never

tilled land (Mid NT; 9 ), or a late successional deciduous forest (Late DF; 0 ).

Figure 2.9 contains the same data represented as a dendogram.

61

 

a Late DF (r1)

 

_. Late DF (r2)

 

 

 

 

 

e Mid NT (r2)

 

 

4 Mid NT (r1)

 

‘-———0 Late DF (r3)

 

 

e Mid HT (r1)

 

Mid HT (r2)
~W Mid HT (r3)

 

 

~-———o Mid NT (r3)

 

_. Early HT (r1)

 

vAg HT (r1)

 

 

-—-—I Early HT (r2)

‘ Early HT (r3)
0-05 I Ag HT (r2)

Figure 2.9. Similarity of methanotroph communities among KBS-LTER treatments. The
Sarenson index was calculated for each pairwise comparison of methanotrophs using
OTUs for all confirmed pmoAs, and then clustered using neighbor-joining with MEGA
(54). Symbols are as follows: Agricultural management of historically tilled land (Ag
HT; V ), early successional plant communities on fields that had been abandoned from
agriculture in 1989 (Early HT; I ), mid-successional plant communities on either
historically tilled land (Mid HT; a. ) or never tilled land (Mid NT; 0 ), or a late
successional deciduous forest (Late DF ; O ).

 

 

 

v Ag HT (r3)

 

 

 

62

References

l.

10.

ll.

Forster P, et al. (2007) Changes in Atmospheric Constituents and in Radiative
Forcing. Climate Change 200 7: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the Intergovernmental Panel
on Climate Change, eds Solomon S, Qin D, Manning M, Chen Z, Marquis M,
Averyt K, Tignor M, & Miller H (Cambridge University Press, Cambridge), pp
129-234.

IPCC (2007) Climate Change 2007: Synthesis Report. Contribution of Working
Groups I, H and III to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. eds Team CW, Pachauri RK, & Reisinger AGeneva,
Switzerland).

Rigby M, et al. (2008) Renewed Growth of Atmospheric Methane. Geophysical
Research Letters 35:doi: 1 0.1029/2008GL036037.

Schlesinger WH & Andrews JA (2000) Soil respiration and the global carbon
cycle. Biogeochemistry 48(1):7-20.

Smith KA, et al. (2000) Oxidation of atmospheric methane in Northern European

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

Grandy AS & Robertson GP (2006) Initial cultivation of a temperate-region soil
immediately accelerates aggregate turnover and C02 and N20 fluxes. Global
Change Biology 12(8): 1 507-1 520.

McDonald IR, Bodrossy L, Chen Y, & Murrell JC (2008) Molecular Ecology
Techniques for the Study of Aerobic Methanotrophs. Appl. Environ. Microbiol.
74(5):1305-1315.

Monson RK, et al. (2006) Winter forest soil respiration controlled by climate and
microbial community composition. Nature 43 9(7077):71 1-714.

Bull ID, Parekh NR, Hall GH, Ineson P, & Evershed RP (2000) Detection and
classification of atmospheric methane oxidizing bacteria in soil. Nature
405(6783):175-l78.

Lau E, Ahmad A, Steudler PA, & Cavanaugh CM (2007) Molecular
characterization of methanotroPhic communities in forest soils. FEMS Microbiol
Ecol 60(3):490-500.

Kuzyakov Y (2006) Sources of CO2 efilux from soil and review of partitioning
methods. Soil Biology & Biochemistry 38(3):425-448.

63

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Hanson PJ, Edwards NT, Garten CT, & Andrews JA (2000) Separating root and
soil microbial contributions to soil respiration: A review of methods and
observations. Biogeochemistry 48(1): 1 15-146.

Cleveland CC, Nemergut DR, Schmidt SK, & Townsend AR (2007) Increases in
soil respiration following labile carbon additions linked to rapid shifts in soil
microbial community composition. Biogeochemistry 82(3):229-240.

Crum JR & Collins HP (1995) KBS Soils. (W. K. Kellogg Biological Station
Long-Term Ecological Research Project, Michigan State University, Hickory
Comers, MI).

Zhou XQ, et al. (2008) Effects of grazing by sheep on the structure of methane-
oxidizing bacterial community of steppe soil. Soil Biology & Biochemistry
40(1):258-26l.

Singh BK, et al. (2007) Effect of afforestation and reforestation of pastures on the

activity and population dynamics of methanotrophic bacteria. Applied and
Environmental Microbiology 73(16):5 1 53-5 161 .

Reay DS, Radajewski S, Murrell JC, McNamara N, & Nedwell DB (2001) Effects
of land-use on the activity and diversity of methane oxidizing bacteria in forest
soils. Soil Biology & Biochemistry 33(12-13):16l3-1623.

Knief C, et al. (2005) Diversity of methanotrophic bacteria in tropical upland soils
under different land uses. Applied and Environmental Microbiology 71(7):3826-
383 l .

Knief C, Lipski A, & Dunfield PF (2003) Diversity and activity of
methanotrophic bacteria in different upland soils. Applied and Environmental
Microbiology 69(11):6703-6714.

Zhou XQ, Wang YF, Huang XZ, Tian IQ, & Hao YB (2008) Effect of grazing
intensities on the activity and community structure of methane-oxidizing bacteria

of grassland soil in Inner Mongolia. Nutrient Cycling in Agroecosystems
80(2):l45-152.

Bomeman J & Triplett EW (1997) Molecular microbial diversity in soils from
eastern Amazonia: Evidence for unusual microorganisms and microbial

population shifts associated with deforestation. Applied and Environmental
Microbiology 63(7):2647-2653.

Allison SD & Treseder KK (2008) Warming and drying suppress microbial

activity and carbon cycling in boreal forest soils. Global Change Biology
14(12):2898-2909.

64

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Schimel JP (1995) Ecosytem Consequences of Microbial Diversity and
Community Structure. Arctic and Alpine Biodiversity: Patterns, Causes and
Ecosystem Consequences, Ecological Studies, eds Chapin 111 FS & Komer C
(Springer-Verlag, Berlin), pp 239-269.

Balser TC & Firestone MK (2005) Linking microbial community composition
and soil processes in a California annual grassland and mixed-conifer forest.
Biageochemistry 73(2):395-415.

Schimel JP & Gulledge J (1998) Microbial community structure and global trace
gases. Global Change Biology 4(7):745-758. ‘

Loreau M & Hector A (2001) Partitioning selection and complementarity in
biodiversity experiments. Nature 412(6842):72-76.

F argione J, et al. (2007) From selection to complementarity: shifts in the causes
of biodiversity—productivity relationships in a long-term biodiversity experiment.
Proceedings of the Royal Society B: Biological Sciences 274(1611):871-876.

Robertson GP, Paul EA, & Harwood RR (2000) Greenhouse Gases in Intensive
Agriculture: Contributions of Individual Gases to the Radiative Forcing of the
Atmosphere. Science 289:1922-1925.

Zhou J, Bruns M, & Tiedje J (1996) DNA recovery fiom soils of diverse
composition. Appl. Environ. Microbiol. 62(2):316-322.

Sambrook J & Russell DW (2001 .) Molecular Cloning: A Laboratory Manual
(Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, USA).

Holmes AJ, Costello A, Lidstrom ME, & Murrell JC (1995) Evidence that
participate methane monooxygenase and ammonia monooxygenase may be
evolutionarily related. FEMS Microbiology Letters l32(3):203-208.

Costello AM & Lidstrom ME (1999) Molecular Characterization of Functional
and Phylogenetic Genes fiom Natural Populations of Methanotrophs in Lake
Sediments. Appl. Environ. Microbiol. 65(11):5066-5074.

Bourne DG, McDonald IR, & Murrell JC (2001) Comparison of pmoA PCR
primer sets as tools for investigating methanotroph diversity in three Danish soils.
Applied and Environmental Microbiology 67(9):3 802-3 809.

Stevenson BS, Eichorst SA, Wertz JT, Schmidt TM, & Breznak JA (2004) New

Strategies for Cultivation and Detection of Previously Uncultured Microbes. Appl.
Environ. Microbiol. 70(8):4748-4755.

65

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

Altschul S, et a1. (1997) Gapped BLAST and PSI-BLAST: a new generation of
protein database search programs. Nucleic Acids Res 25:33 89 - 3402.

Ludwig W, et a1. (2004) ARB: a software environment for sequence data. Nucl.
Acids Res. 32(4):1363-1371.

Konstantinidis KT & Tiedje JM (2004) Trends between gene content and genome
size in prokaryotic species with larger genomes. Proceedings of the National
Academy of Sciences of the United States of America 101(9):3160-3165.

Schloss PD & Handelsman J (2005) Introducing DOTUR, a Computer Program
for Defining Operational Taxonomic Units and Estimating Species Richness.
Appl. Environ. Microbial. 71(3): 1 501-1 506.

Liu WT, Marsh TL, Cheng H, & Forney LJ (1997) Characterization of microbial
diversity by determining terminal restriction fragment length polymorphisms of
genes encoding 16S rRNA. Appl. Environ. Microbiol. 63(11):4516-4522.

Abdo Z, et al. (2006) Statistical methods for characterizing diversity of microbial
communities by analysis of terminal restriction fragment length polymorphisms
of 168 rRNA genes. Environmental Microbiology 8(5):929-938.

Hammer 0, Harper DAT, & Ryan PD (2001) PAST: Paleontological statistics
software package for education and data analysis. Palaeontologia Electranica
4(1):1-9.

Sogin ML, et al. (2006) Microbial diversity in the deep sea and the underexplored
"rare biosphere". Proceedings of the National Academy of Sciences
103(32):12115-12120.

Hanson RS & Hanson TE (1996) Methanotrophic bacteria. Microbiological
Reviews 60(2):439-+.

Stoecker K, et al. (2006) Cohn's Crenothrix is a filamentous methane oxidizer
with an unusual methane monooxygenase. Proceedings of the National Academy
of Sciences of the United States of America 103(7):2363-2367.

Horz H-P, Rich V, Avrahami S, & Bohannan BJM (2005) Methane-Oxidizing
Bacteria in a California Upland Grassland Soil: Diversity and Response to
Simulated Global Change. Appl. Environ. Microbiol. 71(5):2642-2652.

Pacheco-Oliver M, McDonald IR, Groleau D, Murrell JC, & Miguez CB (2002)

Detection of methanotrophs with highly divergent pmoA genes from Arctic soils.
FEMS Microbiology Letters 209(2):313-319.

66

47.

48.

49.

50.

51.

52.

53.

54.

Wertz S, et al. (2006) Maintenance of soil functioning following erosion of
microbial diversity. Environmental Microbiology 8(12):2162-2169.

Enwall K et al. (2007) Long-term impact of fertilization on activity and

composition of bacterial communities and metabolic guilds in agricultural soil.
Soil Biology & Biochemistry 39(1):106-115.

Cavigelli MA & Robertson GP (2000) THE FUNCTIONAL SIGNIFICANCE OF
DENITRIFIER COMMUNITY COMPOSITION IN A TERRESTRIAL
ECOSYSTEM. Ecology 81(5): 1402-1414.

Enwall K, Philippot L, & Hallin S (2005) Activity and composition of the
denitrifying bacterial community respond differently to long-term fertilization.
Applied and Environmental Microbiology 71(12):8335-8343.

Bell T, Newman JA, Silverman BW, Turner SL, & Lilley AK (2005) The
contribution of species richness and composition to bacterial services. Nature
436(7054):1 157-1 160.

Lin JL, et al. (2005) Analysis of methane monooxygenase genes in mono lake
suggests that increased methane oxidation activity may correlate with a change in
methanotroph community structure. Applied and Environmental Microbiology
71(10):6458-6462.

King GM & Nanba K (2008) Distribution of Atmospheric Methane Oxidation and
Methanotrophic Communities on Hawaiian Volcanic Deposits and Soils.
Microbes and Environments 23(4):326-330.

Kumar S, Nei M, Dudley J, & Tamura K (2008) MEGA: A biologist-centric

software for evolutionary analysis of DNA and protein sequences. Brief
Bioinform 9(4):299-306.

67

Chapter 3

The Impact of Long-term Fertilization to the Methanotroph
Communities in Soils

Abstract

Identifying aspects of agricultural management that are causing or contributing to
the loss of methanotroph diversity is crucial if we are to manage lands to conserve or
restore methanotroph diversity, and, in turn, enhance the capacity of soil to serve as a
sink for this potent greenhouse gas. The effect of long-term fertilization, a separate effect
from the short-tenn response to fertilization, is one aspect of row-crop agriculture that is
likely to be a significant disturbance to the methanotroph cOmmunity. We hypothesized
that a consequence of long-term fertilization would be its association with decreases in
methane consumption and methanotroph richness in fertilized forest sub-plots at the
Kellogg Biological Station Long-term Ecological Research Site (KBS LTER). Contrary
to our expectations, we found that long-term fertilization alone did not cause a decrease
in methane consumption nor in methanotroph richness at KBS LTER To determine the
effect of long-term fertilization in other sites we expanded our study by sampling long-
term fertilized agricultural soils from Rothamsted Research, and to include a meta-
analysis of other long-term fertilized soils. Apart from the KBS LTER agricultural soil
we did not find diminished methanotroph richness in long-term fertilized soils, and
comparing between sites, fertilization did not select for similar methanotroph
communities. Rather, methanotroph communities clustered together based on geographic

location regardless of fertilization. The results suggest that each geographic location has

68

a unique methanotroph community, and that it is possible that some of these communities

may be resistant to long-term fertilization.

Introduction

Conversion of well-drained native soils to agricultural use has been demonstrated
to lead to an approximately 70% reduction in the rates of methane consumption
(reviewed by (1), and to cause local extinctions of the methane consuming bacteria
(methanotrophs) who facilitate methane consumption (Chapter 2). Methane is
responsible for 15% of the atmosphere’s total radiative forcing (2), and its atmospheric
levels are currently rising (3). Therefore, identifying the aspects of agricultural
management that are causing or contributing to the loss of methanotroph diversity is
crucial if we are to manage lands to conserve or restore methanotroph diversity, and, in
turn, enhance the capacity of soil to serve as a sink for this potent greenhouse gas.

Common row-crop agricultural practices of tillage, fertilizer, pesticides, and
herbicide application can diminish methane consumption (reviewed in (4)). In addition,
soil pH, water filled pores space, and dry bulk density can directly influence rates of
methane consumption (reviewed by (1, 4, 5). We chose to focus upon fertilization
because ammonia inhibits in situ rates of methane consumption by methanotrophs (6-8)
both through direct competitive inhibition of methane monooxygenase (MMO), and
indirect salt inhibition (9, 10), and has been shown to disturb the methanotroph
community (11, 12). At the Kellogg Biological Station Long-term Ecological Research
site (KBS LTER) Suwanwaree and Robertson (6) found that tillage had no short-term

effect on the rates of methane consumption, but fertilization decreased levels of

69

consumption by as much as 60% in mid and late successional soils. In addition, the
conventional row-crop agricultural soil at KBS LTER, which is regularly fertilized,
harbors few methanotroph species (Chapter 2). We inferred that the persistent long-term
application of ammonia-based fertilizers is a disturbance to the methanotroph community
that is likely to have played a prominent role in causing the loss of methanotroph
diversity from the agricultural treatments at KBS LTER.

There are distinct long and short-term responses to methane consumption rates to
fertilization, and the long-term (chronic) effect is especially pertinent when considering
the effect of fertilization as an agricultural practice on rates of methane oxidation and the
methanotroph community. The short-term response of methane consumption rates after
the application of ammonia-based fertilizer is a dramatic reduction in methane oxidation.
Methane consumption will never fully be eliminated, and there will be a recovery to pre-
fertilization levels of consumption once the ammonia has been consumed as the
competitive and salt inhibition of MO by ammonia is relieved (4, 6, 10). For instance,
Suwanwaree and Robertson (6) found that approximately 7-8 weeks after the application
of fertilizer methane consumption in the late successional deciduous forests (Late DF)
began to recover, and had returned to pre-fertilization levels by 14 weeks. However,
fertilization had no effect on no-till agricultural soil whose rate of methane consumption
remained the same before and afier fertilization. This lack of response to fertilization is
characteristic of the long-term effect of fertilization in which the rate of methane
oxidation is constantly depressed even after new applications of ammonia fertilizer (4, 6,

7).

70

Unlike the short-term response to fertilization, the long-term phenomenon does
not have a straightforward explanation, and suggested reasons for the long-term
attenuation of rates of methane consumption include: fertilization enriching the ammonia-
oxidizing population, who then account for the observed methane oxidation; or
fertilization reducing the methanotroph population. However, no evidence has been
found for ammonia-oxidizers contributing to in situ atmospheric methane consumption,
even after enrichment by fertilization (7, 13, 14). There is evidence that fertilization has
been found to particularly impact Methylocystaceae (11), and reduce methanotroph
abundance (11, 12). Consistent with this trend is the decrease in methanotroph richness
in the long-term fertilized conventional row-crop agricultural treatment (Ag HT) at KBS
LTER (Chapter 2). All of the results suggest that long-term fertilization changes the
methanotroph community, and likely due to those changes, the soils consume less
atmospheric methane.

Therefore, we hypothesized that long-term fertilization would be associated with
decreases in methane consumption and methanotroph richness in fertilized forest sub-
plots at KBS LTER. The hypothesis was tested by determining the methanotroph
richness and community structure with clone libraries of the gene encoding the A subunit
of the particulate methane monooxygenase (pmoA) fiom fertilized sub-plots of the Late
DF treatment at KBS LTER. In addition, the effect of long-term fertilization in other
upland soils was explored through additional pmoA clone libraries from row crop
agriculture and forest soils at Rothamsted Research, and a meta-analysis that included

another long-term fertilized forest (11) and agricultural soil (15) (Table 3.1).

71

Materials and Methods
Site Description and Fertilization history

At KBS LTER, experimental sites were established as 2x2 m plots downhill and
adjacent to each of the three 1 ha replicate plots of the larger late successional deciduous

forest treatment (Late DF) within the KBS LTER. The sites were established in 1995

(16), and the control treatment (Control Sub-plot) has received 0 g N m-2 yr'l. The
fertilized plot (F ertilized sub-plot) sampled in this study, 3N, initially received 3 g N m-2
yr.1 , applied in three 1 g N m'2 applications sometime between April-November. For 4
years (1995-1998), 3 g N m-2 yr'1 was annually applied to the plot as ammonium nitrate
(NH4NO3). No fertilizer was applied during 1999 or 2000. Beginning in 2001, the
annual application was changed to 20—30 g N In2 yr'l, and applied using a backpack
sprayer in two or three 10 g N m.2 applications between April-November as either urea
or NH4NO3. In 2007, only one 10 g N m'2 fertilization took place, as urea, on April 23rd

in replicates l and 2, and on April 24th in replicate 3. Further descriptions of the KBS

LTER and the Late DF sites can be found at http://lter.kbs.msu.edu/.

Rothamsted Research, located in Harpenden, Hertfordshire, United Kingdom,
features the Broadbalk experiment, the longest continuous agricultural experiment in the
world. Since 1843 Broadbalk has been a continuous wheat field, with different plots of

the field receiving different levels of input. Broadbalk plot 8, section 1 (Broadbalk

Wheat), receives 14.4 g N m-2 yr'1 as NH4N03, and was chosen as the long-term

72

fertilized site to be studied due to its exogenous N input being roughly comparable to the
conventional row-crop agricultural soils (Ag HT) at KBS LTER. Two non-fertilized
soils were chosen to compare to the agricultural soil: Broadbalk Wilderness and Knott
Wood. In 1882, one 0.2 ha section of the Broadbalk wheat experiment was fenced and
abandoned from agricultural management. It is presently a wooded area with ash,
sycamore, and hawthorn trees, and is referred to as “Broadbalk Wilderness.” Knott
Wood is at least 300 years old, and is a mixed deciduous woodland. The sites are located
within 1.0 km of each other, and are found on silt to silty clay loams and are Chromic
Luvisols classified as Batcombe series (17). Like the paired sites at KBS LTER, methane
consumption has been shown to be lowest in the Broadbalk Wheat with greater
consumption found in the Broadbalk Wilderness and Knott Wood soils (17, 18). Further
descriptions and details of Broadbalk, Knottt Wood and Rothamsted Research can be
found at http://www.rothamsted.bbsrc.ac.uk/.
Rate Measurements at KBS LTER

At the KBS LTER sub-plots in situ rates of methane cOnsumption were measured
using closed-cover flux chambers (19) from all three replicates of both the control and
fertilized sub-plots. Chambers were constructed with 13 L containers without bottoms
and inserted into the soil to establish approximately 10 L of headspace (20). The
chambers had gas tight lids with rubber septa to allow for gas sampling. Four gas
samples were taken at approximately 15 minute intervals (0, 15, 30, and 45 minutes).
Samples were processed as detailed at http://lter.kbs.msu.edu, and methane
concentrations were determined by GC-FID. Rates were calculated using a best-fit linear

approach. All 4 data points were used to calculate the rate unless the last data point

73

indicated a plateau in consumption, in which case only the first 3 points were used to

calculate the rate. A minority of samples had erroneous points excluded, which were

identified based on a similar outlier also being present in the other trace gases (C02 and
N20) measured fi'om the same sample. Any rate that could not be fit with a linear line

(roughly under an r2 of 0.70) was excluded from the analysis. To determine rate

differences based on treatment, a t-test was performed using the program PAST (21).

Rate measurements were made on June 20th and 27 th, August 1St and 14 th, and

on October 16th in 2007. All rate data were taken 59-177 days afier the last fertilization

event. Suwanwaree and Robertson (6) found that the short-term acute effect of Late DF
fertilization on rates of methane consumption began to abate between 52 and 73 days
after fertilization. Therefore, the comparison of consumption rates is a test of the long-
term fertilization effect, and not the short-term fertilization effect.

Soil Sampling

At KBS LTER, 3 soil cores (2.5 x 10 cm) were collected from the first 2 replicate

sub-plots from the control and fertilized (30 g N m'2 yr'l, 3N) treatments on June 13th

2007. Samples were transported to the laboratory on ice where they were mixed

thoroughly and stored at -80°C until processing.
At Rothamsted Research, 2 soil cores (2.5 x 10 cm) from each site were taken on

February 6th 2008. Each soil core was processed as a replicate as the Rothamsted sites

are not replicated within the greater landscape (http://www.rothamsted.bbsrc.ac.uk/). All

74

samples were air dried for approximately 24hrs prior to sieving (2mm). Once sieved,

samples were stored at -80°C until further processing.

DNA Extraction and pmoA PCR Reactions

DNA was extracted from soil samples with the Mo Bio PowerSoilTM DNA
Isolation Kit, following the manufacturer’s protocol, except that mechanical cell lysis
was performed by bead beating for 45 seconds with KBS LTER soil samples, and bead
beating for 30 seconds with a Fast-Prep FP12 homogenizer (Thermo Scientific, Speed
5.5) with Rothamsted Research soil samples. Three extractions were performed from
each of the Rothamstead Research soil cores.

All soil samples were screened for genes coding for the A subunit of particulate
methane monooxygenase (pmoA) via PCR amplification with the primer sets A189 (5’-
GGNGACTGGGACTTCTGG- 3’) -A682 (5’-GAASGCNGAGAAGAASGC-3’)
(Holmes et al. 1995), and A189-mb661 (5’-CCGGMGCAACGTCYTTACC-3’) (22) in
order to encompass all known pmoA genes (23). No amplification was observed from
A189-mb661. Amplification reactions contained 25 ng of undiluted DNA, 1.25 11.1 1%
BSA, 10 mM dNTPs, 0.2 uM of each primer, 2.5 ul 10x PCR buffer (200mM Tris-HCl

(pH 8.4) and 500 mM KCl), 0.5 1.11 50 mM MgC12 and 1.5 U of T aq DNA polymerase

(Invitrogen, Carlsbad, CA) in a total volume of 25 pl. Reaction conditions were 95 °C for
5 minutes, 15 cycles of a ‘touchdown PCR’ of 95°C for 1 minute, 60°C for 1 minute (-
0.6°C each cycle to 51°C), and 72°C for 1 minute, 1'5 cycles using 51°C as the annealing
temperature, and a final 10 minute extension at 72°C.

Clone Libraries

Cloning was performed with the TOPO-TA cloning kit (Invitrogen, Carlsbad,

75

 

CA) using either vector pCR 4 or pCR 2.1 as per the manufacturer’s protocol.
Transformants were screened via PCR reactions with the primers F2 (5’-
CAGTCACGACGTTGTAAAACGACGGC-3’) and R4 (5’-
CAGGAAACAGCTATGACCATG-3’) (24). One-1.25 ul of the PCR products was

cleaned up via incubation with 0.25 ul of ExoSAP-IT (U sb, Cleveland, OH) for 30

minutes at 37°C. Sequencing was completed at the Research Technology Support

Facility at Michigan State University (RTSF).

Sequences identified by BLASTX (25) as pmoA or the A subunit of ammonia
monoxygenase (amoA), which A189-A682 also amplify, were imported into Arb (26). In
Arb, sequences were translated and aligned using Clustal W. Nucleic acid sequences
were then aligned according to the protein sequence. Sequences from clone libraries were
determined to be the same species if they were 2 94% identical (27) as determined by
DOTUR (average neighbor grouping) (28). pmoA sequences from previously reported
studies were also imported into Arb (26), and aligned as described above. Species were
used to calculate B-diversity (Estimates (http://viceroy.eeb.uconn.edu/EstimateS)) in
order to facilitate the comparison between the methanotroph communities of the different
treatments and studies.

For KBS LTER soils samples, clone libraries were constructed fiom duplicate
PCR reactions from each of 3 soil cores and all 6 reactions were pooled such that there
was only 1 cloning reaction per replicate. For Rothamsted Research samples, clone
libraries were constructed from duplicate PCR reactions of each of the 3 DNA extractions
fi'om the each soil core and, like the KBS LTER samples, all 6 reactions were pooled

such that there was only 1 cloning reaction. Prior to cloning PCR products were digested

76

 

with the PflFI (N eb, Ipswich, MA), and gel extracted with the PrepEase kit (U sb,

Cleveland, OH) to reduce the incidence of cloning amoA and non-specific PCR products.

Results
Rate Measurements

At KBS LTER, average rates of net methane consumption, 21 .5:t4.6 in the control
sub-plot and 18.2:l:2.9 in the fertilized sub-plot, did not significantly differ by treatment
(t=0.58 p=0.57), and long-term fertilization has not caused a decrease in rates of methane
consumption rates (Figure 3.1).

Methanotroph Diversity in KBS LTER Late DF sub-plots

All clone libraries were sampled until rarefaction curves were asymptotic (Figure
3.2a), and due to PCR bias, measures of pmoA abundance were excluded from
comparisons of the methanotroph community (Appendix A).

At KBS LTER, an average of 9.5 methanotroph species was found in the pmoA
libraries from the control sub-plots, while an average of 6.5 was found in the fertilized
sub-plots (Figure 3.1). Despite the higher average richness in the control treatment, the
difference between the treatments was not significant (t=l .18, p=0.44). Between 6-12
methanotroph species were found in the pmoA libraries in this study, all of which fall
within the range of the 6-14 methanotroph species found in previous individual Late DF
libraries constructed from June soil samples (Chapter 2), and not the range of 2-5
methanotroph species found in the individual clone libraries from Ag HT at KBS LTER
(Chapter 2). In addition to the richness differences between the fertilized sub-plots and

Ag HT libraries, only methanotroph Clusters I and II were previously found in Ag HT

77

(Chapter 2), and methanotrophs belonging to clusters 1, II, and Upland Soil Cluster-CC

were found in the fertilized sub-plots (Figure 3.3).

Comparing the methanotroph community composition between the control and
fertilized sub-plots indicated that were some differences between treatments (Figures 3.4a
and 3.5). The control plots clustered together and separately from the fertilized plots
(Figure 3.4a), and the phylogeny of the pmoA sequences recovered from the different
treatments reveal different clustering between the treatments that is being driven by
unshared species in Cluster I and Cluster II as well as richness differences within Cluster
H (Figure 3.5). In Cluster II, 4 methanotroph species were found in the control
treatments, 1 species was found in the fertilized treatments, and no species were shared
between the treatments. Cluster KBSl was also only found in the control plots, adding to
the differences contributing to the clustering of the treatments. Despite the differences
with the control sub-plot, the methanotroph community composition in the fertilized sub-
plots was more similar to methanotroph community in the non-fertilized Late DF
(Chapter 2) than they were to the methanotroph communities previously found in Ag HT
methanotroph clone libraries (Chapter 2) (Figure 3.4b). Additional samples are required
in order to discern, with reasonable confidence, whether the community composition
changes due to long-term fertilization are statistically different.

Methanotroph Diversity at Rothamsted Research

Like the KBS LTER clone libraries, the Rothamsted clone libraries were sampled
until rarefaction curves were asymptotic (Figure 3.2b), and measures of pmoA abundance
were excluded from comparisons of the methanotroph community (Appendix A).

Average methanotroph richness was highest in the agricultural soil, Broadbalk Wheat,

78

.. g...“ fi'

 

 

with an average of 8.5 species, lowest in Knott Wood with an average of 3.5, and in
Broadbalk Wilderness methanotroph richness was 5 (Figure 3.6). In total, 18
methanotroph species were found in Rothamsted soils spread over 5 clusters (Figure 3.7).
Cluster I and Cluster 11 species were found in all three treatments, clusters M90-P96 and

KBSl were only found in Broadbalk Wheat, and Upland Soil Cluster-y was found in

Knott Wood and Broadbalk Wilderness. These community composition differences
drove the separate clustering of Broadbalk Wheat from Knott Wood and Broadbalk
Wilderness (Figure 3.8b).
Comparison of methanotroph communities under long-term fertilization

There was an apparent common effect in the methanotroph communities under
long-term fertilization (Figure 3.6). Lack or replication from all the sites precludes a
robust statistical analyses with the data, but while the Ag HT treatment at KBS LTER has
fewer methanotrophs then Late DF (t=6.4, p=0.02), the decrease in methanotroph
richness in the fertilized sub-plots compared to the control is not significant (see above),
and in all other sites methanotroph diversity increases in the long-term fertilized soils. A
comparison of the methanotroph community composition between the long-term
fertilized and non-fertilized paired soils reveals that the methanotroph communities

cluster according to geographic location (Figure 3.8).

Discussion
The effect of long-term fertilization in Late DF sub-plots at KBS LTER

We found the lack of a difference in methanotroph richness between the control

and long-term fertilized Late DF sub-plots at KBS LTER (Figure 3.1) surprising, and

79

contrary to our hypothesis. The richness difference between Ag HT and Late DF
(Chapter 2), and Suwanwaree and Robertson’s (6) finding that after applying 10 g N In2
of fertilizer to Late DF there was a dramatic, although temporary, decrease in rates of
methane consumption led to the expectation that a decrease in methanotroph richness
would be observed in the fertilized sub-plots in Late DF at KBS LTER The fertilized

sub-plots in Late DF had received 6 consecutive years of fertilization at levels that

'7

exceeded the annual amount of nitrogen applied to Ag HT. In total, 16 10 g N m'2
fertilizer applications were applied to the fertilized forest sub-plots prior the initiation of

this study. The Ag HT plot is a com-soybean-wheat rotation that typically receives 12

 

and 6 g N m'2 in a single application of fertilizer when com and wheat, respectively, are L
grown. Each fertilizer application applied to forest sub-plots was therefore similar to the
amount applied to Ag HT soil, and due to the frequency of applications the annual
amount of nitrogen applied to the forest sub-plots far exceeded the amount of nitrogen
received by Ag HT. Due to the many years of increased fertilizer application we
expected that, like Ag HT (Chapter 2, Figure 3.6), methanotroph richness would decline
in the Late DF fertilized sub-plots.

However, in contrast to our prediction, we found that there was neither attenuated
rates of methane consumption nor methanotroph richness due to fertilization (Figure 3 .1).
The agreement between these two measures is consistent with the previous finding that at
KBS LTER methanotroph richness correlates to rates of methane consumption (Chapter
2). Therefore, without a change in the number of methanotroph species in the Late DF
sub-plots, we would not have expected, nor was there, a change in the rates of methane

consumption. The results are also consistent to the previously observed short-term

80

 

repression of methane consumption in response to fertilization (6) as the measurements of
methane consumption in this study quantified the long-term effect of fertilization, and
care was taken to ensure that methane consumption rates were measured after the short-
term effect of the fertilization had previously been found to subside (6). Thus, while
fertilization of the late successional deciduous forest at KBS LTER is an acute temporary
disturbance to rates of methane consumption (6), its long-term effect was insignificant to
both rates of methane consumption and methanotroph richness.

Our results suggest that the observed decrease in methanotroph richness
associated with agricultural conversion at KBS LTER (between the Ag HT and Late DF
treatments) was not caused by long-term fertilization alone as the methanotroph
community is resistant to the fertilization disturbance, and consequently rates of methane
consumption are resilient to the fertilization disturbance. It is possible that the effects of
fertilizer application are somehow buffered by the Late DF soils such that it does not
represent a disturbance to the methanotroph community, but results fi'om other fertilized
Late DF sub-plots confirm that the methanotrophs would have been exposed to the
ammonia in the fertilizer (Appendix B and (6)), and to have experienced the short-term
inhibition associated with the fertilization (6). Regardless of the mechanism that allows
for the methanotroph community to persist after long-term fertilization the results from
the long-term fertilization in the Late DF sub-plots implicates other changes associated
with agricultural land-use as either causing, or interacting with the short-term fertilization
disturbances to cause the majority of the decrease in methanotroph richness in the Ag HT

treatment at KBS LTER.

81

 

Supporting the possibility that the methanotroph community of the fertilized
subplot is resistant to fertilization is the clustering of its methanotroph communities with
the Late DF methanotroph communities as opposed to the Ag HT communities (Figure
3.4b), the presence of more methanotroph clusters in the fertilized sub-plots then is found
in Ag HT (Figure 3.3), and methanotroph richness measures which are consistent with
those found in Late DF, not Ag HT (Chapter 2). These results suggest that the variation
in methanotroph community composition and methanotroph richness is within the typical
range of variation observed in pmoA libraries from Late DF, and not the range of
variation seen in Ag HT pmoA libraries. The trend of all sub-plot libraries to be similar
to Late DF data indicates that the community has likely not changed after long-term
fertilization and is possibly resistant to the fertilization disturbance.

The efiect of long-term fertilization on methanotroph richness in other soils
The response of the methanotroph community to only long-term fertilization
alone has only been studied in one other well-drained late successional forest soils (7,

11). Those studies took place at the Harvard Forest where pine and hardwood soils have
been fertilized with 15 g N m'2 yr.l since 1988, and the methanotroph community in the
fertilized treatments has changes despite no apparent change of methanotroph richness
(Figure 3.6): Gulledge et al. (7), after 10 years of fertilization, found a lowered Km for

methane oxidation in fertilized treatments — indicative of a methanotroph community that
had changed to one less competitively inhibited by ammonium; and after 12 years of
fertilization methanotroph abundance had declined, with Methylocystaceae abundance

being most affected (1 1).

82

It is possible that additional years of fertilization would yield discernible
methanotroph community changes, as the differences in the methanotroph community at
the Harvard Forest were found well after 6 years of consecutive fertilization - when we
assessed the methanotroph community at KBS LTER. However, Castro et a1. (29) did
find that the rate of methane consumption had declined 15-64% due to fertilization in the
Harvard forest soils after 6 years of consecutive fertilization. The different effect on
methane consumption rates between 6 year fertilized soils at KBS LTER and Harvard
Forest suggests that differences between the sites, either physical properties or the
endemic methanotroph and microbial communities are likely responsible for the
contrasting results.

Further implicating differences in methanotroph communities as part of the reason
for the contrasting results of long-term fertilization between KBS LTER and Harvard
Forest, is the absence of a typical effect to methanotroph richness in other long-fertilized
soils (Figure 3.6), and the clustering of methanotroph communities according to
geographic location (Figures 3.8). Comparing between sites, fertilization did not select
for similar methanotroph communities; rather, methanotroph communities clustered
together based on geographic location regardless of whether they were or were not
fertilized. Therefore, each geographic location appears to harbor a unique assemblage of
methanotrophs, and the contrasting responses between soils of the methanotroph
community to long-term fertilization and land use change is probably at least partially
due to fundamentally different methanotroph communities. Further confirming the
biogeography of methanotrophs is a similar comparison of methanotroph community

composition in a variety of upland soils (Appendix C).

83

With inherently different methanotroph communities at each site it would not be
surprising if the response of the community to long term fertilization and/or land-use
were unique for each geographic location. For instance, just using rates of methane
consumption to determine the impact to the methanotroph community, distinct
methanotroph communities were probably responsible for the different responses to long-

term fertilization observed between birch and spruce taiga forests (30). In addition,

 

F

Nyerges and Stein (31) found a broad range of sensitivities to ammonia inhibition in four is

methanotroph isolates, suggesting methanotroph communities of differing composition

will react distinctively in response to ammonia. %
In addition to the unique methanotroph community, unique physical and E,

environmental characteristics from each geographic location are also probably playing a
role in dictating the response of the methanotroph community to changing land use. For
instance, if we compare the long-term fertilized forest at the Harvard Forest to the long-
term fertilized Late DF sub-plots at KBS LTER, we find a number of differences that
might be contributing to the observed differences in their methanotroph communities to
long-term fertilization. At Harvard Forest, the forest soils are stony to sandy loams
(Entic Haplorthods of the Gloucester series) (32), have low pH (3.0-4.0) (7), have a
history of agricultural use (33), high nitrogen retention after fertilizer application (nearly
100%) (32, 34), and, especially in the hardwood forest, low rates of nitrification (33, 34).
The absence of nitrification for at least the first 6 years of fertilization in the hardwood
forest (34) and the nitrogen retention of the soils (32, 34) are signs that the methanotroph
community was exposed to fertilization disturbances far longer then the methanotrophs in

the fertilized sub-plots at KBS LTER. At KBS LTER fertilized sub-plots, there is robust

84

nitrification in the Late DF sub-plots ((6) and Appendix B), and while the methanotroph
community is exposed to the fertilizer, the ammonia is consumed within approximately
two months at which point rates of methane consumption begin to recover.

Therefore, the nitrification rates at KBS LTER may buffer the methanotroph
community from the fertilization disturbance enough to make them resistant, and
methane consumption rates resilient, to the fertilization disturbance. This possibility
could be experimentally tested by inhibiting nitrification after fertilization and seeing if
the methanotroph community and methane consumption rates remain unaffected by the
fertilization disturbance. If the methanotroph community were found to be resistant to a
fertilization disturbance and not just buffered by increased nitrification - then those would
be methanotrophs that would be candidates to try to get to colonize agricultural soils.
We expect fertilization resistant methanotrophs to likely be more successful in increasing
the rates of methane consumption in agricultural soils as they would be able to persist in
the face of one of the disturbances associated with agricultural practices. Conversely, if
the methanotroph community is buffered from the fertilization disturbance by increased
nitrification, then increasing nitrification in agricultural soils may provide an avenue
towards increased atmospheric methane consumption in agricultural soils.

Conclusion

We found that long-term fertilization alone did not cause a decrease in methane
consumption nor in methanotroph richness at KBS LTER. Comparing those results to
other long-term fertilized soils found that apart from the KBS LTER agricultural soil we
did not find diminished methanotroph richness in long-term fertilized soils, and

comparing between geographic sites, fertilization did not select for similar methanotroph

85

H.11-

 

 

communities. Rather, methanotroph communities clustered together based on geographic
location regardless of fertilization. The results suggest that each geographic location has
a unique methanotroph community, and that it is possible that some of these
methanotroph communities, including the one in the KBS-LTER fertilized sub-plot, may
be resistant to long-term fertilization. The results also indicate that managing land to
preserve or restore methanotrophs in agricultural soils at KBS LTER and elsewhere will
have to take into account many management factors, not just fertilization, and have to
understand the interaction of each factor with each site’s unique methanotroph and

microbial communities and soil properties.

Acknowledgements
All methane rate measurements from KBS LTER and were taken by Neville Millar, who
also helped with the calculations of methane flux. I’m also grateful for the help I

received at Rothamsted Research from Penny Hirsch, Ian Clark, and Paul Poulton.

86

r. ..s. enAJK..rnu-A . .Ll. . . k

.828 :8 88:3. 8 282 828:3”— q
.83on 228823508 2 me 30:22.. owabfi 5. 2:3 AN .2220. 3.282: .3636... e we $286.5 2.. 32822013
.86on 228.825.: wN .8 9.02:0... 09326 .8 2:? AN 289:—UV 3.282: 3223?: v we 9:883 05 928235

328:9... a: _ o. 2&2 8.8223— .

2 : 88 .3 3 =3 3 322m 8...... 5... 683.23.. .55 32c... .35..
2 .. SON .3 3 =3 3 322m .23: no“. 325“. 2.: mm: 32c“. .35..
2 : NSN .3 3 :3 N 8.3 32.35 .25 .328. 2.... am: 329. .233:
.62... 2.: oN 322m 32...: N ..3. 8.3.3.3 82.232 .35 23 33.5 822m 32222.... m32.3.
2.2.3 3...... 3. 322m .323: . ..3. .2923 32.35“. .35 23 33.5 822m 323.22 32.3.
2.2... 2.: .N 222m .23: N 3.3. 8.923 .238 .35 23 3...: 5.33 322.35 32.3.
2.22.. 2.: 8 222m .23: . ..3. 8.923 .238 .35 23 3.5.. 8.33 323.22 mm2.3.
.3: 38 .3 3 .25. N 23.3,: .2358 8.33 35253.3... .223
Am 3 Sea ..a .o .225— c 35:35. 22.32.33 2.83... So .8» @— 288.83— =o_§m ..acoEtoN—xm 3.8.—mm
G: meow ..a .o .225— m 28:22. .88"— :ooumugm catfim ESEEaxm Eco—am
>23 v.2... 2 28323. 3:23 N and .325» £3395 «£8383. 33:22.3—
»uBm m2... 9. 80352 @225 _ mom 48:3 xfinvam «29.8.9.3— Baegzom
beam 3:. mm Eovmfiv. @833 N mom $02.62.? ..72—Roi 4283QO 338.353—
xcam m3... 9 Eocwcg @825 _ mom $2525., 2.332685 45.883. 35:853—
xvam m2... .3... Eovwfiv— 28:5 N com .383 :20— ..22383- 38:82qu
zuBm m2: mm EonwEv. 32:5 _ and 683 .35— ..28383. 368.353—
N .2926 2. 322m .23: N ..3. .35 23 33.5 8.33 32222.. 32.3.
N .2936 8.. 8.2m .2225 M. as. .3... 23 3.25 822m 323.2... 32.3.
N 323.30 N3 835 32.35 N 3.3. H: w... 3.3.. 822m 322.33. m3.2.3.
N .236 wNN 322m .223: N. as. .2. w... 3...: 822m 323.22 mm2.3.
3 a a 3% mzlm
3.403%

.333... mi. 5 33: 823.5: WEEK 28 82m 05 mo ban—8.5 ._.m 03$.

87

Net Methane Consumption Methanotroph Species
(g CH4-C ha"1 day‘)

 

 

 

 

 

 

 

 

 

 

 

28- —14
24— -12
20— I -10
16j - 8
T
12- - 6
8 - - 4
4 - — 2
o _
Control Sub-plot Fertilized Sub-plot Control Sub-plot Fertilized Sub-plot

Figure 3.1. Differences between Late DF fertilized and control treatments at KBS LTER
in average net methane consumption and the average number of methanotroph species
(defined as pmoA sequences having 94% average nucleotide sequence similarity).
Fertilization has no effect on either measure (t=0.58, p=0.57 for consumption; and t=1.18,
p=0.36 for richness). Measures for net methane consumption are averages from 3
replicates of each treatment, while the methanotroph species are averages from 2
replicates of each treatment. Net methane consumption from the same 2 replicates as
those with methanotroph species data yields the same result (20.9243 and 19.2233
average net methane consumption of the control and fertilized sub-plots, respectively,
and t=0.28 p=0.79). Error bars represent standard error.

88

P

Number of Methanotroph Species

Number of Methanotroph Species

 

 

 

4o -
__ Fertilized Sub—plot 1

35 4 — — - Fertilized Sub-plot 2
""""" Control Sub-plot 1

30 + ' ' - - Control Sub-plot 2

25 _

20 _

15 -

1o - __________________________________

5 .. ,5 1,5". g ’: 2’ ——————————————
/"
o I I I I T 1 I
Number of Sequences
40 _
Broadbalk Wheat 1

35 d ---------- Broadbalk Wheat 2
--------- Broadbalk VWdemess 1

30 d W Broadbalk VWdemecs 2
“WW Knott Wood 1

25 ‘ ------------ ~— Knott Wood 2

20 -

 

 

 

 

Number of Sequences

Figure 3.2. (A) Rarefaction curves from pmoA clone libraries constructed from the
control and fertilized sub-plots of the late successional forest at KBS LTER. Libraries
were constructed from DNA extracted from soil sampled in June 2007. (B) Rarefaction
curves from pmoA clone libraries constructed from Rothamsted Reasearch soils sampled
in Feburary 2008. All curves were constructed using data fiom neighbor joining matrixes
from Arb (Ludwig et al. 2004), with curves calculated by DOTUR (Schloss and
Handelsman 2005). Methanotroph species are defined as pmoA sequences having 94%
average nucleotide sequence similarity. Error bars representing 95% confidence intervals
were omitted for the sake of clarity.

89

 

 

a—proteobacteria

Upland Soil
. I Clustera ’N
l Type II

 

 

 

 

 

methanotrophs
y-proteobacteria
{ amoA
y-proteobacten'a Type I Verrucomicrobia
methanotrophs} pmo A
K Upland Soi
Clustery
MR1. I
f O I Cluster ll :RA21 Unknown
Classification
Unknown
Classification M90-P96
\ O KBS1
B-proteobacteria 4
GMOA Cmnothn‘x polyspora) y-proteobacteria
Presumably
0.10 Crenarachaeota Clusterl . '- a—proteobacteria
_ amoA

 

Figure 3.3. Phylogenetic tree of selected partial pmoA and amoA protein sequences fi'om
public databases and translated from PCR-based clone libraries fiom late successional
deciduous forest sub-plots. The tree is based on 164 amino acid positions using Phylip
Protein Maximum Likelihood as implemented in ARB (26). Bolded clades were
recovered from the late successional forest sub-plots. The symbols adjacent to the clade
reflect the clade’s recovery fiom either the control treatment (0 ) or the fertilized
treatment (I). The scale bar represents 10 PAM units.

90

 

Control Subplot 1

 

 

 

Control Subplot 2

 

Fertilized Subplot1

 

 

 

Fertilized Subplot 2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.05
B. __{——Late DF Rep 1, June '05,60°C-51°C
Control Subplot 1
Late DF Rep 1, June '05, 62°C-56°C
Control Subplot 2
Fertilized Subplot 1
Late DF Rep 1, June '06, 60°051°C
Late DF Rep 3, June '06, 60°C-51°C
Fertilized Subplot 2
—- l LateDFRep1,June'O5,51°C
' LateDFRep1,June'05,48°C
!" Ag HT Rep 1, June '06, 60°C-51°C
l——- Ag HT Rep 2, June '06, 60°C-51°C
Lae DF Rep 1, Deoerrber'04,62°C-56°C
Ag HT Rep 1, June'05,48°C
I Ag HT Rep 1, June '05,62°C-56°C
1 AgHTRep1,oewrper'o-1,62°056°c
l——l
0.05

Figure 3.4. (A) Similarity of methanotroph communities among fertilized and control
sub-plot replicates. (B) Similarity of methanotroph communities between fertilized and
control sub-plot replicates with all other pmoA libraries from KBS LTER Expect for the
fertilized and control sub-plots, the pmoA clone libraries were first reported in Chapter 2,
and details of their construction can be found there. Both dendograms are based on
Serenson index calculations for each pairwise comparison of the methanotroph
communities using pmoA species (defined as pmoA sequences having 94% average
nucleotide sequence similarity), and then clustered using neighbor-joining with MEGA
(3 5). The scale bars represent at 0.05 change in the Sorenson index.

91

 

 

Figure 3.5. Neighbor joining phylogenetic tree of the partial nucleic acid sequences of
pmoA and amoA from reference sequences, and KBS-LTER clone libraries from the
control and fertilized sub-plots soil sampled in June 2007. The tree is rooted with the
branching from the amoA sequences fiom Nitrosomonas eutropha, Nitrosococcus
mobilis, and Nitrosovibrio tenuis. Nodes representing the 17 different methanotroph
species (defined as pmoA sequences having 94% average nucleotide sequence similarity)
found in late successional deciduous forest sub-plots are circled. The symbols in the
circles reflect species’ recovery from either the control treatment (0) or the fertilized
treatment (I ).

92

 

 

 

 

%W——l:l) Crenothrix polyspora

—I:7) B-proteobacteria amoA

)Type I methanotrophs y— proteobacteria

Nitrosococcus oceani strain AFCZ4P amoA (AF509002)
Nitrosococcus oceani strain SW amoA (AF509003) ) Y— —pr0teObaCteria amOA

)T ell methanotro hs, oc— roteobacteria

Uncultured bacterium MR4 (AF200727)
Fertilized sub—plot 2, 0708 CO7

Fertilized sub-plot 1, Fert BOS -

Control sub— —plot 2, 0708 G04 . I Upland SOII CIUSter 0‘4

Control sub- -plot 2, Control H02

Uncultured bacterium Hold :1 (AF148523)

Control sub— —plot 1 1CDF-

Uncultured bacterium DGGE band5 E5FB- f (AJ579663)
Uncultured bacterium clone CL10 (A J699 O7)

 
  
 
 
 
 

    
 
  
   
    
  
  

 

 

   

 

Uncultured bacterium DGGE band CF- h (3AJ868282)
cControl sub- plot 1,1—CDF 1 CO 02 .
gontr ol sub- plot 2, 0708 E05

Control sub- -plot 2, 0708 G06 .

     

 

Cluster |l

Control sub- lot 1, 1CDF B10

  

 

    

0 j
Fertilized su plot 2 0708 COB
Fertilized su -plot 2, 0808 H02 I]

Control sub- -p|ot1, 1CDF- 1 804
Control sub- plot 1, C1DF- 1 805

—E%%t‘i%.it%%& 1’ 18% 19302 0 j)KBs1
M90 P69

{ Uncultured bacterium DGGEb and E56F- -a (A J579662)
Uncultured bacterium MR1 (AF200729)

Control sub- plot 1,1CDF-1 A08

Control sub-plot 2, Control C12 . I MR1
Fertilized sub plot 1,0708 A02

Fertilized sub— plot 1,0708 A01

Uncultured bacterium amo/pmoPN_ O- 5#2 (EF452672)

)RA21

)Verrucomicorbia pmoA
Uncultured bacterium clone B7233- 79 (AJ()564439)313

 
 
      

 

 

 

v

 
       

      
   
    
 
 
     
 

 

 

Uncultured bacterium clone pmoA— U2- D0295
Fertilized sub- p—Iot 2, 0708 C03 I
Fertilized sub- plot 2 0708 B03
Uncultured bacterium DGGE6 band E33b- -a (AJ579661)
Control s-pub ot,1 1CD FD
O I

Fertilized subl— —ptlot’1, 13x1 E02
Fertilized sub- plot 1 1208 A02
Control sub- plot 1 1CDF— 18 06
Control sub- plot 1,1CDF CO4 L
_ Methanotroph K3 16 (AF547178)
—- Uncultured bacterium clone pLWPmoA- -24 (DQ067064)
Uncultured bacterium clone LOPA12 4 (AF358041)
Uncultured bacterium clone CL27 (AJ699310)
Fertilized sub-plot 1, Fert 807
Control sub- plot 2, Control E07 . I
Fertilized sub- -p|ot 1 13x1 A04
Control sub- —p|ot2, Control 805 Cluster I
Uncultured bacterium clone CL49 (AJ699315)
Fertilized sub— plot 1 1208 (302 !
Methanotroph K1 -8 (AF547175)
Methanotroph K2- 14 (AF54718707)
Methanotr orph K3— 21 (AF 547
Uncultured bacterium clone pmoAa4 (D0367742)
Uncultured bacterium c|c[>3n19 CL5 (AJ699316)
Control sub- -p|ot1, 1CDF . I
Fertilized sub- plot 2, Fert E04
Control sub- p—Iot 2, 0708 F02
Uncultured bacterium DGGE band L58-b2 (AJ868244)
Fertilized sub- plot 2, 0708 D07 .1
Fertilized sub- plot 2 0708 D06
Control sub- plot 1 tCDF- 1 CO7
Fertilized sub- -plot 2, Fert F05 . I
Fertilized sub- plot 2, 0708 COB
Control sub- -p|ot 2, Control E01
Control sub- plot 1, iCDF B12 . J
Control sub-plot 1, 1CDF Bot

    
   
  
 
 
 
  
  
    
 
  
   
  

 

 

 

 

Figure 3.5

0.10

93

 

 

 

14
7 .Non-fertilized Soils

I Fertilized Soils
12 4

10-i

Methanotroph Species

 

Figure 3.6. Methanotroph richness in paired sites featuring long-term fertilized (black
shading) and non-fertilized soils (grey shading). The three sites to the left compare
fertilized sites in the context of agricultural management, while the two sites on the right
are fertilized forests whose only land management is fertilization. Error bars represent
standard errors as the averages are reported for sites that featured replicated sites.
Additional site details are provided in Table 3.1.

94

 

 

cat-proteobacteria

 

 

 

Upland Soil JR1
Cluster a
Type II
/ methanotrophs
y-proteobaoteria
amoA f
Type I r-
methanotrophs Verrucomicrobia -
_ \ pmoA
y—proteobactena Upland Soil
Cluster y
1 Wood I
1Wildemess .5
r
0 “we“ MR1
Cluster II L,
1 Wood RA21
4 Wheat - -
M90-P96 Classification
Unknown 0 Wood
Classification KBS1 OWilderness
0 Wood 1 Wheat
0.1 0 Wilderness
— 1 Wheat
p—proteobaoten‘a
amoA ‘ Crenothrix polyspora) y-proteobacteria
‘ Presumably
Crenaracheeota Cluster I a—proteobacteria
amoA 3 Wood
4 Wildemess
7 Wheat

Figure 3.7. Phylogenetic tree of selected partial pmoA and amoA protein sequences fi'om
public databases and translated from PCR-based clone libraries fi'om late successional
deciduous forest sub-plots. The tree is based on 164 amino acid positions using Phylip
Protein Maximum Likelihood as implemented in ARB (26). Bolded clades were
recovered from Rothamsted Research soils. Beneath each bolded clade is a listing of the
number of species within that clade found in which Rothamsted Research treatment:
Knott Wood (Wood), Broadbalk Wilderness (Wilderness) and Broadbalk Wheat (Wheat).
The scale bar represents 10 PAM units.

 

95

0.241
0.18 -

0.12-

      
   
 

Coord‘nate 2

 

 

O ' Sakerat
Experimental
-0.06 . Station
-0. 12 -
127
-0.18 -
-0.24 -
-0 3

 

30.48 oi -o.é2 -o.é4 -o.1'6 ode d 0.68 0.16 0.524
Coord'nate1

B. Broadbalk Wildemess 2
HE m w... 2
Knott Wood 1

H— Broadbalk Wildemess 1
Broadbalk Wheat 11:}

Broadbalk Wheat 2 o

Fertilized Sub-plot 2i} \‘

 

 

Rothamsted Research

 

 
 
 
  
  
 
   

 

Fertilized Sub-plot 11?
Late DF 3 KBS LTER

Control Sub-plot 2
Control Sub-plot 1
Late DF 1

Evergreen Forest .
___r—i:—Reforested Plantation ) 322%? ”mm”
Comfieldfi
Pine Forest, Control
—‘i:—r_l=ine Forest, Fertilizedfi ) Harvard Forest
l—-t Hardwood Forest, Fertilizedtf}

0.1

 

 

 

Figure 3.8. Similarity of methanotroph communities in paired sites featuring long-term
fertilized and non-fertilized soils. (A) The Sarenson index was calculated for each
pairwise comparison of methanotroph species using two-dimensional non-metric
multidimensional scaling (21). (B) The same data as in (A), but clustered using neighbor-

joining with MEGA (35) and displayed as a dendogram. Labels with a 7:? are sites that
have been fertilized long-term. Additional site information is provided in Table 3.1.

96

 

 

References

l.

10.

11.

12.

Smith KA, et al. (2000) Oxidation of atmospheric methane in Northern European
soils, comparison with other ecosystems, and uncertainties in the global terrestrial
sink. Global Change Biology 6(7):791-803.

IPCC (2007) Climate Change 2007: Synthesis Report. Contribution of Working
Groups I, II and IH to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. eds Team CW, Pachauri RK, & Reisinger AGeneva,
Switzerland).

Rigby M, et al. (2008) Renewed Growth of Atmospheric Methane. Geophysical
Research Letters 35:doi:10.1029/2008GLO36037.

Hutsch BW (2001) Methane oxidation in non-flooded soils as affected by crop
production - invited paper. European Journal of Agronomy 14( 4):237- 260.

King GM (1997) Responses of atmospheric methane consumption by soils to
global climate change. Global Change Biology 3(4):351-362.

Suwanwaree P & Robertson GP (2005) Methane Oxidation in Forest,
Successional, and No-till Agricultural Ecosystems: Effects of Nitrogen and Soil
Disturbance. Soil Science of America Journal 69: 1722-1779.

Gulledge J, Hrywna Y, Cavanaugh C, & Steudler PA (2004) Effects of long-term
nitrogen fertilization on the uptake kinetics of atmospheric methane in temperate
forest soils. FEMS Microbiology Ecology 49(3):389-400.

Chan ASK, Steudler PA, Bowden RD, Gulledge J, & Cavanaugh CM (2005)
Consequences of nitrogen fertilization on soil methane consumption in a
productive temperate deciduous forest. Biology and Fertility of Soils 41(3):182-
1 89.

Conrad R (1996) Soil microorganisms as controllers of atmospheric trace gases
(HZ, CO, CH4, OCS, N20, and NO). Microbial. Rev. 60(4):609-640.

Gulledge J & Schimel JP (1998) Low-Concentration Kinetics of Atmospheric
CH4 Oxidation in Soil and Mechanism of NH4+ Inhibition. Appl. Environ.
Microbial. 64(11):429l-4298.

Lau E, Ahmad A, Steudler PA, & Cavanaugh CM (2007) Molecular
characterization of methanotrophic communities in forest soils. FEMS Microbial
Ecol 60(3):490-500.

Maxfield PJ, Hornibrook ERC, & Evershed RP (2008) Acute impact of

agriculture on high-affinity methanotrophic bacterial populations. Environmental
Microbiology 10(7): 191 7-1924.

97

 

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Kolb S, Knief C, Dunfield PF, & Conrad R (2005) Abundance and activity of
uncultured methanotrophic bacteria involved in the consumption of atmospheric
methane in two forest soils. Environmental Microbiology 7(8):1150-1161.

Carini SA, Orcutt BN, & Joye SB (2003) Interactions between methane oxidation
and nitrification in coastal sediments. Geamicrabialagy Journal 20(4):355-3 74.

Knief C, et al. (2005) Diversity of methanotrophic bacteria in tropical upland soils

under different land uses. Applied and Environmental Microbiology 71(7):3826—
3 83 l .

Ambus P & Robertson GP (2006) The effect of increased 11 deposition on nitrous
oxide, methane and carbon dioxide fluxes from unmanaged forest and grassland
communities in Michigan. Biogeochemistry 79(3):315-337.

Willison TW, Oflaherty MS, Tlustos P, Goulding KWT, & Powlson DS (1997)
Variations in microbial populations in soils with different methane uptake rates.
pp 85-90.

Hutsch BW, Webster CP, & Powlson DS (1994) METHANE OXIDATION IN
SOIL AS AFFECTED BY LAND-USE, SOIL-PH AND N-FERTILIZATION.
Soil Biology & Biochemistry 26(12):]613-1622.

Robertson GP, Paul EA, & Harwood RR (2000) Greenhouse Gases in Intensive
Agriculture: Contributions of Individual Gases to the Radiative Forcing of the
Atmosphere. Science 289:1922—1925.

Millar N & Kahmark K (2008) Field Gas Sampling Protocol (Bucket Method).
(W. K. Kellogg Biological Station Long-Tenn Ecological Research Project,
Michigan State University, Hickory Comers, Nfl).

Hammer 0, Harper DAT, & Ryan PD (2001) PAST: Paleontological statistics
software package for education and data analysis. Palaeontalagia Electronica
4(1):l-9. -

Costello AM & Lidstrom ME (1999) Molecular Characterization of Functional
and Phylogenetic Genes from Natural Populations of Methanotrophs in Lake
Sediments. Appl. Environ. Microbial. 65(11):5066-5074.

Bourne DG, McDonald IR, & Murrell J C (2001) Comparison of pmoA PCR
primer sets as tools for investigating methanotroph diversity in three Danish soils.
Applied and Environmental Microbiology 67(9):3802-3 809.

Stevenson BS, Eichorst SA, Wertz JT, Schmidt TM, & Breznak JA (2004) New

Strategies for Cultivation and Detection of Previously Uncultured Microbes. Appl.
Environ. Microbiol. 70(8):4748-4755.

98

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Altschul S, et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of
protein database search programs. Nucleic Acids Res 25:3389 - 3402.

Ludwig W, et al. (2004) ARB: a software environment for sequence data. Nucl.
Acids Res. 32(4):]363-1371.

Konstantinidis KT & Tiedje JM (2004) Trends between gene content and genome
size in prokaryotic species with larger genomes. Proceedings of the National
Academy of Sciences of the United States of America 101(9):3160-3165.

Schloss PD & Handelsman J (2005) Introducing DOTUR, a Computer Program
for Defining Operational Taxonomic Units and Estimating Species Richness.
Appl. Environ. Microbiol. 71(3):]501-1506.

Castro MS, Steudler PA, Melillo JM, Aber JD, & Bowden RD (1995) Factors
Controlling Atmospheric Methane Consumption by Temperate Forest Soils.
Global Biageachemical Cycles 9:1-10.

Gulledge J, Doyle AP, & Schimel JP (1997) Different NH4+-inhibition patterns
of soil CH4 consumption: A result of distinct CH4-oxidizer populations across
sites? Soil Biology and Biochemistry 29(1): 1 3-21.

Nyerges G & Stein LY (2009) Ammonia cometabolism and product inhibition
vary considerably among species of methanotrophic bacteria. FEMS
Microbiology Letters 297(1): 1 3 1-136.

Aber JD, Alison M, Boone R, Melillo JM, & Steudler P (1993) Plant and Soil
Responses to Chronic Nitrogen Additions at the Harvard Forest, Massachusetts.
Ecological Applications 3(1):156 - 166.

Compton JE, Watrud LS, Porteous LA, & DeGrood S (2004) Response of soil
microbial biomass and community composition to chronic nitrogen additions at

Harvard forest. F arest Ecology and Management 196(1): 143-158.

Aber J, et al. (1998) Nitrogen Saturation in Temperate Forest Ecosystems.
BioScience 48(11):921-934.

Kumar S, Nei M, Dudley J, & Tamura K (2008) MEGA: A biologist-centric

software for evolutionary analysis of DNA and protein sequences. Brief
Biainform 9(4):299-306.

99

Chapter 4

Conclusions and Future directions

In this thesis there are multiple lines of evidence suggesting that multiple aspects
of the methanotrophs’ habitat are disrupted by row-crop agriculture at KBS LTER: (a)
The decline of both methane consumption and methanotroph diversity in row-crop
agricultural soils; (b) the 75 years required after the cessation of agricultural land
management for methane consumption rates and methanotroph diversity at KBS LTER to
achieve the current rate of methane consumption and diversity of the native soils
(Chapters 2 and 3). An array of row-crop agriculture related factors, in addition to, and
perhaps in interaction with long-fertilization that was investigated in Chapter 3, may all
be disturbing the methanotroph community and are discussed below. Together, the net
effect of row-crop agriculture at KBS LTER is the apparent destruction of methanotroph
niches and subsequent loss of methanotroph richness and rates of methane consumption.

Each of the row-crop agriculture related factors could be explored in the future, as
our ability to potentially manage lands to conserve or restore methanotroph diversity, and
enhance the capacity of the methane soil sink, will rely on understanding the effect of
these potentially pivotal variables on the methanotroph community. Only 1 of the pmoA
phylogenetic clusters found at KBS LTER has not been reported in other upland soils
(Chapter 2), and we recovered that cluster in the Broadbalk wheat soil at Rothamsted
Research (Chapter 3). Therefore, while each sites’ methanotroph community
composition is unique (Chapter 3, Appendix C), at least some of the methanotroph
species are going to be closely related to the methanotrophs that we find at KBS LTER.

The discovery of the effects of land management practices to specific methanotroph

100

species would likely be applicable to other sites where the same or closely related species
are present.

Previous studies have found that the common row-crop agricultural practices of
tillage, fertilizer, pesticides, and herbicide application can negatively impact methane
consumption (reviewed in (1)). These practices are all featured in KBS LTER’s
conventional row-crop agricultural treatment, and therefore may at least be partially
causing the 7-fold land-use related decrease in methane consumption and methanotroph
diversity observed in Ag HT (Chapter 2). In addition, land use related changes to soil
properties like pH, water filled pores space, and dry bulk density may be contributing to
the decline of methane consumption and the methanotroph community in Ag HT. Each
has been shown to directly influence rates of methane consumption (reviewed by (1 -3)),
and changes in the methanotroph community have been found to be associated with
different pHs (4, 5), changes in the successional stage of the plant community (Chapter 2
and (5, 6)), temperatures, and precipitation levels (7). None of these factors were
investigated in the present study as we chose to only investigate long-term fertilization.

The short-term effect of tillage on methane consumption at KBS LTER was
negligible (8), and led to the assumption that its long-term effect would be minor.
However, reduced tillage has been shown increase methane consumption (reviewed by
(1), and the disturbance to soil structure as a result of tillage is potentially a contributor to
the 75-100 years it takes to for methane consumption rates and methanotroph richness to
recover fi'om agricultural land use (Chapter 2 and (3)), and with long-term fertilization’s
minor effects, tillage’s long-term effect on methanotroph communities merits further

investigation.

101

For instance, tillage may be changing the soil structure such that dry bulk density
and water filled pore spaces change and cause less methane to be available to the
methanotroph community. Increases in both dry bulk density and water filled pore
spaces limit gas diffusion, and have been correlated to decreases in rates of methane
consumption (2, 3). Dry bulk density at KBS LTER has only occasionally been
measured previously at KBS LTER (http://lter.kbs.msu.edu), and therefore we cannot
determine if either dry bulk density or water filled pore spaces (the determination of
water filled pore spaces depends on the measure of dry bulk density) are affecting
methane consumption.

Pesticide and herbicide application, other agricultural practices at KBS-LTER,
have also been shown to negatively impact the rate of methane consumption at other sites
(1, 9-11), and can alter the methano/methylotroph community (9). The magnitude of the
chemical impact on methane consumption and the methanotroph community can greatly
vary depending upon chemical and soil type (1, 9, 10, 12). For instance, long term
application of atrazine and metolachlor, two herbicides among those used at KBS-LTER,
has been found to not cause a difference in the rate of methane consumption, and to only
cause minor changes in the composition of the methano/methylotroph community (10).
Also, a study contrasting the effect herbicides and fertilization found that the
methano/methylotroph community clustered according to the type of fertilization, and
that methane consumption rates did not significantly decline due to herbicide treatment
(12). The conventional agricultural soil at KBS LTER has been treated with a variety of
chemicals (http://lter.kbs.msu.edu), so a definitive determination of herbicide or

pesticide effects on the KBS LTER methanotroph community will be difficult. In

102

addition, the organic agricultural soil at KBS LTER despite having no herbicide or
pesticide application since 1989 still has a low rates of methane oxidation (13); indicating
that the influence of herbicide and pesticide application on rate and the methanotroph
community is likely to minimal. However, herbicide or pesticide application may
nevertheless be influencing the loss of methanotroph richness at KBS LTER.

Decreasing pH has also been correlated to reduced methane consumption
(reviewed by (3 )), but the reverse trend has been observed at KBS LTER. Neither rates
of methane consumption nor methanotroph richness decline with pH at KBS LTER. The
Late DF soils have the lowest pH (approximately 5.3, http://lter.kbs.msu.edu), but have
the most methanotroph richness and the greatest rates of methane consumption, while the
Ag HT soils have the highest pH (approximately 6.2) with the least methanotroph
richness and the lowest rates of methane consumption. Notably, while declining pH is
not affecting methanotroph richness at KBS LTER, it may be limiting methanotmph
richness in the Rothamsted Research soils. There, Knott Wood has the least
methanotroph richness, and the lowest pH (4.0) while Broadbalk wheat has the highest
pH (6.7) as well as the most methanotroph richness (14, 15).

Future directions into determining the cause behind the decrease in methanotroph
diversity and methane consumption associated with row-crop agriculture at KBS LTER
might best be focused upon tillage, changes in soil structure, changes associated with the
plant community (discussed in Chapter 2), and the possible interaction of these factors
with fertilization and pesticide and herbicide application. Each of these directions would

shed light on the niche for KBS LTER methanotrophs, and lead to insights that could

103

 

predict and explain the response of methanotrophs and, most importantly, rate of methane

consumption to management changes at KBS LTER.

104

 

References

1.

10.

11.

Hutsch BW (2001) Methane oxidation in non-flooded soils as affected by crop
production - invited paper. European Journal of Agronomy 14( 4):237- 260.

King GM (1997) Responses of atmospheric methane consumption by soils to
global climate change. Global Change Biology 3(4):351-362.

Smith KA, et al. (2000) Oxidation of atmospheric methane in Northern European
soils, comparison with other ecosystems, and uncertainties in the global terrestrial
sink. Global Change Biology 6(7):791-803.

Knief C, Lipski A, & Dunfield PF (2003) Diversity and activity of
methanotrophic bacteria in different upland soils. Applied and Environmental
Microbiology 69(11):6703-6714.

Knief C, et al. (2005) Diversity of methanotrophic bacteria in tropical upland soils
under different land uses. Applied and Environmental Microbiology 71(7):3826-
383 1 .

Zhou XQ, et al. (2008) Effects of grazing by sheep on the structure of methane-

oxidizing bacterial community of steppe soil. Soil Biology & Biochemistry
40(1):258-261.

Horz H-P, Rich V, Avrahami S, & Bohannan BJM (2005) Methane-Oxidizing
Bacteria in a California Upland Grassland Soil: Diversity and Response to
Simulated Global Change. Appl. Environ. Microbiol. 71(5):2642-2652.

Suwanwaree P & Robertson GP (2005) Methane Oxidation in Forest,
Successional, and No-till Agricultural Ecosystems: Effects of Nitrogen and Soil
Disturbance. Soil Science of America Journal 69:1722-1779.

Mertens B, Boon N, & Verstraete W (2005) Stereospecific effect of
hexachlorocyclohexane on activity and structure of soil methanotrophic
communities. Environmental Microbiology 7(5):660-669.

Seghers D, et al. (2003) Effect of long-term herbicide applications on the bacterial
community structure and function in an agricultural soil. FEMS Microbiology
Ecology 46(2): 139-146.

Seghers D, et al. (2003) Pollution induced community tolerance (PICT) and
analysis of 16S rRNA genes to evaluate the long-term effects of herbicides on
methanotrophic communities in soil. European Journal of Soil Science 54(4):679-

684.

105

 

12.

13.

14.

15.

Seghers D, Siciliano SD, Top EM, & Verstraete W (2005) Combined effect of
fertilizer and herbicide applications on the abundance, community structure and
performance of the soil methanotrophic community. Soil Biology & Biochemistry
37(2):l87-193.

Robertson GP, Paul EA, & Harwood RR (2000) Greenhouse Gases in Intensive
Agriculture: Contributions of Individual Gases to the Radiative Forcing of the
Atmosphere. Science 289:1922-1925.

Goulding KWT, Willison TW, Webster CP, & Powlson DS (1996) Methane
fluxes in aerobic soils. Environmental Monitoring and Assessment 42(1): 175-187.

Hutsch BW, Webster CP, & Powlson DS (1994) METHANE OXIDATION IN
SOIL AS AFFECTED BY LAND-USE, SOIL-PH AND N-FERTILIZATTON.
Soil Biology & Biochemistry 26(12): 1 61 3-1622.

106

 

Appendix A
Assessment of PCR bias

Introduction

To determine if relative abundance, in addition to richness, could be used in
determining and comparing the methanotroph community structure, the bias from the
polymerase chain reaction (PCR) that amplified pmoA in the tRF LP analysis and some of
the clone libraries was assessed. PCR bias is the over-amplification of specific templates
which results in the post-amplification concentrations of those templates being much
greater then their pre-amplification concentrations (1, 2). If there is PCR bias then the
abundance measures fiom PCR based community analyses are misleading, erroneous
conclusions could be made regarding the dominant and rare organisms in a given
community, and measures (i.e. diversity indices) that rely on relative abundance will be
skewed.

The methanotroph community analyses in this thesis (Chapters 2—3) utilize the
A189-A682 primer pair (3) to amplify pmoA. Any PCR bias associated with the primer
pair has not been quantified, nor has the consistency of the output from the primers and
the specific PCR conditions been addressed previously. Thus, PCR bias was assessed
with tRF LP profiles whose initial templates were defined artificial communities of pmoA
and amoA (The A subunit of the ammonia monooxygenase (amoA) is also amplified by
A189-A682) that varied the concentration of initial template. Ideally, if there is PCR
bias it will be consistent throughout the mixtures so that despite the bias the relative

abundances measures could be included in the community analyses of the pmoA

107

 

 

community. While not suitable for quantitative comparisons, the measures would be the

same regardless of concentration changes within the mix, justifying their usage.

Methods

The template for amplification reactions was 4 different mixes of purified
plasmids, and each of the purified plasmids. The plasmids contained pmoA PCR
products fiom soils at the Kellogg Biological Station Long Term Ecological Research site
TOPO TA cloned into vector pCR 4 or pCR 2.1 (Invitrogen, Carlsbad, CA) (Chapter 2).
The plasmids represented 6 pmoA species and 1 species of the A subunit of the ammonia
monooxygenase (amoA), which A189-A682 also amplifies. The plasmids were mixed
such that the concentrations of all the species were held approximately constant except
for one pmoA species, Cluster I A (Table 1). Each mix was run as a tRF LP either 2 or 3
times with each tRF LP beginning with new PCR amplifications.

The PCR reactions contained a total of 30 pg of template, 1.25 pl 1% BSA, 10
mM dNTPs, 0.2 pM of each primer (A682 was labeled with fluorophore 6-
carboxyfluorescein (6-Fam)), 2.5 pl 10x PCR buffer (200mM Tris-HCl (pH 8.4) and 500
mM KCl), 0.5 pl 50 mM MgClz and 1.5 U of Taq DNA polymerase (Invitrogen,

Carlsbad, CA) in a total volume of 25 pl. Reaction conditions were 95 °C for 5 minutes,
15 cycles of a ‘touchdown PCR’ of 95°C for 1 minute, 62°C for 1 minute (-O.4°C each
cycle to 56°C), and 72°C for 1 minute, 15 cycles using 56°C as the annealing
temperature, and a final 10 minute extension at 72°C.

For each tRF LP, 4 replicate PCR reactions were pooled. PCR products were

purified using a MinElute column (Qiagen, Valencia, CA). In a 50 pl reaction, 300ng of

108

 

purified product was digested with 1.5 U Tau I (F ermentas, Glen Burnie, MD) by

incubating at 55°C for 1 hr 30 min. To inactivate the enzyme the DNA was precipitated

as follows: The sample was diluted to 500 pl, followed by the addition of 50 pl 3M
sodium acetate, 1 pl or 2.5 pl 10 mg/ml glycogen, and 500 pl isopropanol, and holding
on ice for at least 5 minutes. The DNA was then pelleted by centrifugation at 16,000 x g
for either 5 (1 pl glycogen) or 10 minutes (2.5 pl glycogen). The supernatant was
decantated, and the pellet washed with 500p] of 80% ethanol followed by centrifugation
for 2 minutes at 16,000 x g, and removal of the supernatant. After a 30 second
centrifugation additional ethanol was removed, and the DNA was air dried for 5-10
minutes before resuspension in 20 pl of water. In an 18-22 pl reaction 140 ng of DNA

was digested with 2.5 U SspI (Neb, Ipswich, MA) in a 1hr 30min incubation at 37°C.
After heat inactivation at 65°C for 20 minutes 6 U of BstUI (Neb, Ipswich, MA) was
added, incubated at 60°C for 1hr 30min, and inactivated by adding 0.8 pl of 0.5 mM

EDTA. Capillary electrophoresis of the tRF LP reactions was then performed with a 5 fu
cutoff at RTSF.

Individual peaks were distinguished fiom the background signal and binned using
TRFLP-Stats (4). In TRFLP-Stats default settings were used except for the standard
deviation cutoff, which was increased to 4.5. The resulting cutoff was approximately 25
fu. The tRF LP profiles of the individual plasmids ensured that the specific bin belonging
to each species was correctly identified. Any other peaks were excluded from the
analysis, and the relative abundance of the remaining peaks were normalized such that

the absolute abundance of every treatment was 100. The amount of relative abundance

109

 

excluded from any one tRF LP profile before normalization was between 0-11% with all

tRFLPs except one having <6% excluded.

Results

The pmoA species Cluster I B was consistently over-amplified in every mix, while
species Cluster I A was over-amplified except for when its initial abundance was less
then 10% of the initial artificial community (Table 1). The over-amplification of Cluster
I A and Cluster I B species was dramatic, and in some mixes was approximately double
their expected output (Figure 1, Mixes 2-4). In addition, when Cluster I A’s
concentration in the artificial community was increased, its relative abundance in the
output also increased. However, the increase was not proportional, and increasing initial
concentrations of Cluster I A led to increased over-amplification. For instance, between
Mix 2 and Mix 4, the concentration of Cluster 1 A in the output was expected to increase
2.3 fold, but its actual output increased 2.9 fold.

Cluster II A was the only species whose actual output was consistently close to its
expected output, although when Cluster I A was expected to be less then 10% of the
output (Mix 1), it too was over-amplified (Table 1). pmoA species Cluster II B was also
overamplified in Mix 1 but otherwise, it, along with pmoA speices Cluster I C, Cluster II
D, and amoA species A, was under-amplified (Figure 1, Table 1). Cluster I C had the
lowest amount of starting template in all the mixes, and its under-amplification resulted
in the species not being recovered in two mixes (Mix 1 and 4), and amoA species A was

also not recovered in Mix 4.

110

 

Discussion
The cause of the observed PCR bias is unclear. It is possible that it is at least
partially caused by the different binding energies from the degenerate positions within the

primers (1). The GC content differences from the degenerate positions can result in a
2°C difference in annealing temperature, and subsequently a greater proportion of GC-

rich template can bind to the primers and cause over-amplification as compared to
primers with A or T at those positions (1). Due to the ambiguous positions there are 16
template sequences that a reverse primer could bind, and 4 sequences that a forward
primer could bind. Only 2 of the species in this experiment (species Cluster II B and
Cluster II D) actually have identical primer binding sites. As a result, the likelihood of
the primer binding sites having different binding energies is high, and is possibly
contributing to the PCR bias.

However, if the degenerate primers were the primary cause of the bias we would
expect to see (1) all GC-rich primer binding sites over-amplified, and (2) consistency in
the pattern of PCR bias as the same species should be always be over-or under-amplified.
We do not see either. With the exception of the consistent over-amplification of species
Cluster I B, the results for all species were inconsistent; with some species being under-
amplified in some mixes and over-amplified in others (i.e. Cluster II B). In addition,
Cluster I B only has A or Ts in degenerate positions, and not G or Cs. We therefore
conclude that the primer pair A189-A682 is not the primary cause of the PCR bias, and
its biases are likely no worse than most other primer pairs.

Regardless of the underlying cause, the inconsistent pattern of over- and under-

amplification argues for the exclusion of the relative abundance data. Supporting that

111

 

conclusion is that as the concentration of the Cluster I A species increased with each mix,
its over-amplification became worse. This is disconcerting because it indicates that
communities that are seemingly dominated by an over-amplified species are more likely
to have erroneous measures of relative abundance. The over-amplified Cluster I A
dominates the tRF LP profiles fiom Ag HT (Chapter 2), and therefore the relative
abundance measures from the Ag HT methanotroph communities are those that are most
likely to be erroneous due to the PCR bias.

For instance, the first replicate of Ag HT had Cluster I A at 98% relative
abundance, and Cluster II A and MRI were both at 1% relative abundance. Looking at
that result, or if put into a diversity index (i.e Simpson or Shannon), the presence of
Cluster II A and MRI would be discounted due to Cluster I A’s dominance of the
community. However, the PCR bias indicates that the 98% relative abundance is likely
over-amplified at least 2-fold, so the true relative abundance of Cluster 1A is probably
closer to 50%. If Cluster II A and MR1 are being under-amplified, as our results
indicate for Cluster II A (see Mixes 3 and 4; MRI was not included in this experiment),
then their true relative abundances are much higher then 1%. Therefore, the fairest and
most representative way to present the data is to just consider whether a species is present
or absent.

Despite species Cluster I C not being recovered in two mixes, its lack of recovery
was not associated with either high or low species Cluster I A input; indicating that the
loss of its presence may have been independent of the amount of bias in a reaction. Mix
1, with the least over-amplification, did not recover cluster I C, and neither did Mix 4,

which had the greatest over-amplification of species Cluster I A. Communities with

112

 

highly over-amplified species are expected to have an increased chance of not being able
to recover methanotroph OTUs (as shown by the amoA A species being recovered in all
mixes but Mix 4), but the inability to recover Cluster I C in the least biased mix suggests
that the recovery of methanotroph OTU or species that are in low abundance is variable,
and just as possible in communities with limited PCR bias and those with considerable
PCR bias. Considering that 5 tRF LP profiles were summed to represent one replicate
(Chapter 2), the expectation of recovering a low-abundance species even in treatments
with the highest over-amplification of species Cluster I A is reasonable.

Therefore, due to the PCR bias and the presence of nearly all methanotrophs in
each mix, all methanotroph community analyses are presented only with
presence/absence data (richness, Sorenson index), and exclude the relative measures of
methanotroph abundance. The over-amplification of pmoA species Cluster I A and
Cluster I B, and the considerable deviation of all species from their expected output led to
the conclusion that abundance measures should not be included in the tRF LP or clone
library community analyses of the pmoA community. If future investigators can identify
more reproducible amplification conditions, then the use of the relative abundance

measures could be justified and include in the analyses of the methanotroph community.

113

 

 

 

 

 

 

 

 

 

o 0 NH m.m < <23
H H NH mgr. a HH .336 <23
H 4 NH m.m m HH .386 <23
o 0 NH m.m < HH .336 <23 v .5...
o o N H.N u H.330 <23
o 2 NH m.m m H.330 <23
H. 3 mm 3: < H.336 <23
m m 4H NH. < <23
4 H. «H N... a HH .336 <23
o 4 4H NH. m HH .336 <23
N 2 4H N... < HH .336 <23 n x...
N N m m.N u H.336 <23
H H 4H NH. m H.335 <23
S 3 HN 3 < H .336 <23
N N 2 3. < <23
m m H 0.4 0 HH .386 <23
H N H a... a HH .336 <23
4 NH mH 2v < HH .336 33 N x...
N N m N.N u H.325 <23
N NH. 3 3. m H.336 <23
m N H o... < H .336 <23
N H. 2 a... < <23
N S 2 we a HH .335 <23
H NN 3 as. a HH .336 <23
H N 3 a... < HH .325 <23 H x...
o o 3 m.N u H.336 <23
H on 3 o... m H.325 <23
H m m m.N < H .325 <23
.95 35 33.6 .253 05 3 15 2.9.6 388.5 2: 33 gonna! 2: 5 (an 3.8%
taste-um 00:23:30 250.0: 009.25 he 8:23:50 9583. 38.—Ea... 3 252.3 (OE. .3 (OER

360% «SEN. new «SEQ .8 33:38:26 .8595 cacao“. .8 ”59:0 .253 can 593 uofioaxo .coEmanoU .H.< use.

.Nwoarmmgx :3 ..mEtQ 9.: no mm... «UN. 05 “Woman 3 flow:

114

I Expected
50 '1

Observed

 

Relative Abundance in tRFLP Output (% of Total)

 

 

Cluster IA Cluster I B Cluster I C Cluster H A Cluster H B Cluster 1] D amoAA

pmoA or amoA Species

Figure A.1. Expected and actual relative abundance of the artificial pmoA and amoA
community Mix 2.

115

References

1.

P012 MF 8:. Cavanaugh CM (1998) Bias in Template-to-Product Ratios in
Multitemplate PCR. Appl. Environ. Microbiol. 64(10):3724-37 30.

Suzuki M & Giovannoni S (1996) Bias caused by template annealing in the
amplification of mixtures of 16S rRNA genes by PCR. Appl. Environ. Microbiol.
62(2):625-630.

Holmes A], Costello A, Lidstrom ME, & Murrell IC (1995) Evidence that
participate methane monooxygenase and ammonia monooxygenase may be
evolutionarily related. FEMS Microbiology Letters 132(3):203-208.

Abdo Z, et al. (2006) Statistical methods for characterizing diversity of
microbial communities by analysis of terminal restriction fragment length
polymorphisms of 16S rRNA genes. Environmental Microbiology 8(5):929-
938.

116

 

 

Appendix B

Ammonia and Nitrate Before and After Fertilization
in a KBS LTER Fertilized Sub-plot

Introduction

In addition to sampling one of the KBS LTER Late DF fertilized sub-plots, we
also established a 10x Fertilized sub-plot. The 10x sub-plot was designed to allow
investigation into the short-term effect of fertilization on the methanotroph community,
but its community was not investigated after we found that long-term fertilization had no
effect on methanotroph richness (Chapter 3). However, the results from the nutrient
analyses of the inorganic N from the 10x sub-plots from before and after fertilization is
presented here to illustrate that the nitrogen in the applied urea fertilizer is converted to
ammonia that can be recovered from the soil. Even though we have no direct evidence
that the ammonia from the fertilization inhibits the activity from the 1 OX sub-plot
methanotrophs, we can confirm that the ammonia concentration after urea fertilization in
the soil is high enough to be predicted to inhibit the rate of methane consumption. In
addition, the data is also provided to illustrate the turnover of ammonia to nitrate that is

indicative of considerable nitrification in the 10x sub-plot soils.

Methods

A 2x2 m sub-plot was established in each Late DF replicate adjacent to the other

fertilized sub-plots. The 10x sub-plot received one application of 33 g N m'2 yr'l as
urea, which was applied using a backpack sprayer on June 5th 2007. Three soil cores (2.5

x 10 cm) were collected from all 3 replicates on the following dates: June 1St 2007, June

117

13th 2007, June 27th 2007, August 14th 2007, and October 17th 2007, representing 4 days

before fertilization, and 8, 22, 47 and 111 days after fertilization.

Approximately 10g of each soil core was extracted with 100ml of 1 M KC], and
prepared for analysis on a Flow Injector Analyzer as per the protocol available at
http://lter.kbs.msu.edu/ (Soil Inorganic N). The three measures fiom each plot were then
averaged to be a composite measure for that replicate, and the average of three biological

replicates is represented in the Figure B. 1.

Results and Discussion
In the 10x sub-plot, 8 days afier fertilization, a big spike in ammonia

concentrations was observed as the enzyme urease quickly converted urea to ammonia (l ,

2). The concentration of ammonia at that date was 8 NH4-N ppm (or 8 pg NH4-N g soil-

1). That concentration, based on modeling by Hiitsch (3) is predicted to inhibit methane

consumption. By 22 days after fertilization, nitrate concentrations have begun to rise as
the ammonia is niuified, and by the third measurement, 47 days after fertilization, nitrate
has peaked, and concentrations of ammonia have fallen. This conversion of most of the
ammonia to nitrate in 47 days after fertilization is evidence of robust nitrification, and
probably lessens the effect of the fertilization disturbance upon the methanotroph
community as exposure to ammonia may be relatively limited. Therefore, we conclude
that the urea fertilizer was converted rapidly to ammonia, the concentrations of ammonia
were high enough to have inhibited methane consmnption, and ammonia was quickly

nitrified to nitrate.

118

 

Acknowledgments
Neville Millar was an equal partner in the soil extractions, and was responsible for

running the samples through the Flow Injector Analyzer.

119

I l
i 10.0 '1 r 5.0 i
E +Control ‘ i

l

   
  
   
     

 

i

‘ 9.0 3‘ NH4-N ‘ 4.5
l
t 8.0 ‘ +10x ‘ 4.0
i Fertilized ‘
‘ 7.0 i NH4'N l? 3.5
i ‘ +Control \

6.0 4 N03“ r 3.0 E
l 3. ‘ a
i a. +10x

i

(Z, 5‘0 \ Fertilized t 2'5 z.“
‘t :5? 1 NO3-N ; O
)2 4.0 ‘ » 2.0 Z
. l
l 1 .
l 3.0 1 - 1.5
' l
l 2.0 t t 1.0 I
l
l l .
3 1.0 ~ 1 0.5
l - .

0.0 w e . . 3&1 0.0

-5 20 45 70 95 120
Days Since Fertilization

 

Figure B. 1. Ammonia and nitrate concentrations in fertilized and control Late DF sub-
plots at KBS LTER. Error bars represent the standard error.

120

References

1.

Malhi SS, Grant CA, Johnston AM, & Gill KS (2001) Nitrogen fertilization
management for no-till cereal production in the Canadian Great Plains: a review.
Soil and Tillage Research 60(3-4):101-122.

Rochette P, et al. (2009) Banding of Urea Increased Ammonia Volatilization in a
Dry Acidic Soil. J Environ Qual 38(4):]383-1390.

Hutsch BW (2001) Methane oxidation in non-flooded soils as affected by crop
production - invited paper. European Journal of Agronomy 14( 4):237- 260.

121

 

Appendix C
Biogeography of Methanotrophs in Well Drained Soils

Introduction

To confirm and extend the biogeography observed in the long-term fertilized sites
in Chapter 3 (Figure 3.8), and to determine the biogeography of methanotrophs in well
drained soils, a meta-analysis was performed using 27 clone libraries of the A subunit of
the particulate methane monooxygenase (pmoA) (Table C.1). pmoA is found in all
known methanotrophs (1) except two Methylocella strains (2). Due to the near ubiquity
of pmoA, and the inference of function fiom the presence of the gene, it was chosen over

comparing libraries that assess the methanotroph community via 16S ribosomal genes.

Methods

Using only pmoA sequences allowed for all of the sequences to be imported into
Arb (3) where they were translated and aligned using Clustal W. Nucleic acid sequences
were then aligned according to the protein sequence. Sequences from clone libraries were
determined to be the same species if they were _>_ 94% identical (4) as determined by
DOTUR (average neighbor grouping) (5). In order to facilitate the comparison between
the methanotroph communities of the different soils species were used to calculate fi-
diversity with the Sorenson index (PAST, (6) and Estimates
(http://viceroy.eeb.uconn.edu/EstimateS)). The clustering between the communities was
visualized using a neighbor-joining dendogram.

Only pmoA libraries fiom well drained soils that consume atmospheric methane

were included in the analysis. Libraries from environments that are net sources of

122

atmospheric methane like landfill cover soils, mine soil, rice paddies, wetlands, etc, were
excluded fiom the analysis. If a study fi'om a well drained soil included a library from an
experimental treatment of the same land use type (e.g. increased atmospheric C02; (7))
that library was excluded, and only the pmoA libraries from the control treatments were
included. Only pmoA libraries that had at least 10 sequences were included in order to
ensure that under-sampled methanotroph communities would not confound the analysis.
Doing so excluded many pmoA libraries that were based on results from denaturing
gradient gel electrophoresis (i.e. (8, 9)), and the Harvard forest and Sakerat Experimental
Stations clone libraries included in the biogeography analysis of long-term fertilized sites
(Chapter 3).

Included in the final analysis were 27 pmoA libraries from well drained soils from
Germany, the United Kingdom, the United States, and New Zealand (Table C. 1). The
libraries were from 9 studies (Chapter 2, Chapter 3, (7, 10-14)) , and included a total of

1,560 pmoA sequences from a variety of forest, pasture, shrub land and agriculture soils.

Results and Discussion

The meta-analysis of methanotroph community compositions in well drained soils
(Figure C. 1) revealed a distinct biogeography, and confirmed the patterns seen in Chapter
3 with long-term fertilized sites. Methanotroph communities generally clustered
according to geographic location. The two exceptions to that pattern was 1 library from
a German forest clustering with the Rothamsted Research libraries, and the two Hawaii

soils not sharing any species, and therefore clustering separately. The Hawaii forests are

123

on soils that vary in age by approximately 50,000 years so it is not surprising that those

libraries did not conform to the pattern of biogeography evident elsewhere (Table C.1).

124

:5 ”SN 352 2... wee.
:5 ”SN £52 H...e 35.
E «SN .3 a no:

2: HSN .He He Nam
a: HSN .He .0 N8...
2: HSN ._e a Ne3H
Hm: HSN .He .0 N3...
8: EN 2.; HesW .Hmem
: HH NSN .He .0 swam
HH HH 88 .He .0 .HweHm
HH : 82 ._e .0 swam
HH HH 88 .He .0 35m
2 HH EN .3 He .35
2: EN .H.w He :35
HNHH 38 He He 35.

H .326
N .236
m .3920
m .0325
m .3925
m .2920
N .225
m 8320
N .2325
N .8325
N .325

N 5525
g

.20 38» can boHaEHxER—a m. 3.28 on... N

55: 96. 2o .8» occdnA a co v2.82 2 86.8 2t. e
.cowebi v5 62.8.9095 938883 .50 .3 £0.6— EoBE< a

6.23 :8 28:92 2 .52 332.30% H.

.88on snoboHHafioE o no maggot 09.26 SH .23 Am bag—UV 355: 3.6365 0 .Ho 9:883 05 85838:
.on9on £32859:— wd .Ho maggot omflo>a SH .23 AN BEE—0:01.25: .3239: v .8 9:85;». 05 358505

mNN
moocozmom <oEm
.Ho 53:52

8.3 use:
moaaum flog—a:
8.3m 3...:

e535. 3...:
.535 3...:
scenes. 3...:
5835. 3:5

van—aoN >62
23.3 302
3:3 3.2
22.3 .32
BanN 302
233 302

van—uoN 302
.3250

5835. use:
5839 uses
5830. 3...:
.835. 3:5
5835. 3...:
Scenes. use:
83% 93:5
8.3 3...:
8.3 Bee:
8.3 Bee:
8.3 Bee:
8.5. 8%:
33d

N368.— Eceuv. @822
e698"— unEeguE @222
mace—mafio

33m 05.. 38m
33m 333m .3332

2.3m HBO
25m but?
.83“— :ooom 530er20
838..

22385
3.38;
25H
22mg
2.:
.88...

N as. $2.3 H323
H 33. .325, vHHaHHceeaH
N as. 38:32.3 3333
H as. 30:52.3 3323
N .3 683 .55.
H .3. 683 .85.

N e3. .235 .380 an 23
H e3. 83.25 .228 ..E 23
N .5. .an 23
MH e3. .3 23
N e... H: w<
NH .3 H: m<
3%

323:3. HauHmo—oB a: H 2 Hugo. 328:3.”— H

3c: «>2 wov— 2:82
33H 8220 ..c 525 .2339» nose—U.
anti 5%....
5:35
E320
.5530
.5520
cuss— E33380
8:033
.5533
32333
32383
3:55
3:5.—
33... 59580

4:288”— 38:553—
..noHaomox “c2255.:—
vnufimox vonEasHox
«£888.: BHmEafiox
«£859.3— vonEafio:
4.3.383— 33:853—

...HmHHH 8:3 .8323. 32.3.

Rm: 8:3 .3323 32.3.

Rm: 8:3 22335 32.3.

:35 8:3 .8365 MW3:3.

Rm: 8:3 Hansen 33:3.

Rm: 8:3 HeeHweHeHm 32.3.

am

$95 25 5 com: 8:95: V2.3 can 8H5 2: .8 >355 HQ 035...

125

 

Broadbalk Wilderness 2
Knott Wood 2

 

Knott Wood 1
Gottingen Forest Rothamsted Research

 

 

 

Broadbalk Wilderness 1

 

 

Broadbalk Wheat 1

 

 

 

Broadbalk Wheat 2

 

 

 

 

Late DF Control Sub-plot 1
—|—— Late DF Control Sub-plot 2

 

 

 

 

 

 

 

Late DF 1
Late DF 3 KBS LTER

 

Oak Stand
__: Spruce Stand

 

 

 

 

Acacia Forest, Mauna Kea, Hawaii

Alder Stand Glsburn

 

Pine Stand

 

 

 

 

Jasper Ridge Grassland
Metrosideros Forest, Kilauea Volcano, Hawaii

 

Westview Pine

 

 

Old-growth Beech Forest
Waiouru Pasture

 

 

 

t

 

 

l——l
0.1

 

Westview Pasture New Zealand
Waiouru Shrubland
Puruki Pine
Puruki Pasture

Figure C.1. Similarity of methanotroph communities in soils from around the globe. The
dendogram is based on Sarenson index calculations for each pairwise comparison of the
methanotroph communities using pmoA species (defined as pmoA sequences having 94%
average nucleotide sequence similarity), and then clustered using neighbor-joining with
MEGA (48). The scale bars represent at 0.1 change in the Sorenson index. Additional site
information is provided in Table C. 1.

126

 

References

1.

10.

ll.

Kitmitto A, Myronova N, Basu P, & Dalton H (2005) Characterization and
structural analysis of an active particulate methane monooxygenase trimer from
Methylococcus capsulatus (Bath). Biochemistry 44(33):]0954-10965.

Dunfield PF, Khmelenina VN, Suzina NE, Trotsenko YA, & Dedysh SN (2003)
Methylocella silvestris sp nov., a novel methanotroph isolated from an acidic

forest cambisol. International Journal of Systematic and Evolutionary
Microbiology 53:1231-1239.

Ludwig W, et al. (2004) ARB: a software environment for sequence data. Nucl.
Acids Res. 32(4):]363-1371.

Konstantinidis KT & Tiedje JM (2004) Trends between gene content and genome

size in prokaryotic species with larger genomes. Proceedings of the National
Academy of Sciences of the United States of America 101(9):3160-3165.

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

Hammer 0, Harper DAT, & Ryan PD (2001) PAST: Paleontological statistics
software package for education and data analysis. Palaeontologia Electronica
4(1): 1-9.

Horz H-P, Rich V, Avrahami S, & Bohannan BJM (2005) Methane-Oxidizing
Bacteria in a California Upland Grassland Soil: Diversity and Response to
Simulated Global Change. Appl. Environ. Microbiol. 71(5):2642-2652.

Knief C, Lipski A, & Dunfield PF (2003) Diversity and activity of
methanotrophic bacteria in different upland soils. Applied and Environmental
Microbiology 69(11):6703-6714.

Zhou XQ, et al. (2008) Effects of grazing by sheep on the structure of methane-
oxidizing bacterial community of steppe soil. Soil Biology & Biochemistry
40(1):258-26l.

Singh BK & Tate K (2007) Biochemical and molecular characterization of
methanotrophs in soil from a pristine New Zealand beech forest. FEMS
Microbiology Letters 275(1):89-97.

Singh BK, et al. (2007) Effect of afforestation and reforestation of pastures on the

activity and population dynamics of methanotrophic bacteria. Applied and
Environmental Microbiology 73(16):51 53-5 161 .

127

 

12.

13.

14.

Ricke P, Kolb S, & Braker G (2005) Application of a newly developed ARB
software-integrated tool for in silico terminal restriction fragment length

polymorphism analysis reveals the dominance of a novel pmoA cluster in a forest
soil. Applied and Environmental Microbiology 71 (3): 1 671-1673.

Reay DS, Radajewski S, Murrell J C, McNamara N, & Nedwell DB (2001) Effects
of land-use on the activity and diversity of methane oxidizing bacteria in forest
soils. Soil Biology & Biochemistry 33(12—13): 1613-1623.

King GM & Nanba K (2008) Distribution of Atmospheric Methane Oxidation and

Methanotrophic Communities on Hawaiian Volcanic Deposits and Soils.
Microbes and Environments 23(4):326-330.

128