. 11¢ JN
< » :1 ‘ Pf:
hum? e . 92

3..

mm;

A: a
2.0!
4.35.: J _

‘ v , b. . rum».
.fT. Wmtﬁm «wavy mph
a, {$3M he

.. .umw. :

HL

‘3! .
(ll 1:! I
.426 15.1» .3
4:) . . . 1.“..-

.il

.._ .3»

ﬁx: .5

._
a s .. ‘

_ , .131:
Ema um. L 5&3.

._ .awaﬁ.

:

3 V...)

.v
x!-

 

.,.
.: . i: 1..
‘31.. .._

 

.Eéﬁmg 63......

.

 

THCFS

IHIHIUIHHIIUIIUHHHII’HHNHIIHIIHIIHHI

93 01770 0539

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

ECOSYSTEM CONSEQUENCES AND SPATIAL DISTRIBUTION
OF SOIL MICROBIAL COMMUNITY STRUCTURE

presented by

Michel Andre Cavigelli

has been accepted towards fulfillment
of the requirements for
Dept. of Crop and Soil
Ph.D. degree in Sciences
Ecology and Evolutionary
Biology and Behavior
W. K. Kellogg Biological
Station

(QMM

Ma jOl‘ professor

 

 

Date low???

 

MSU i: an Afﬁrmative Action/Equal Opportunity Institution O~ 12771

 

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 I DATE DUE I DATE DUE

0717 0°.
JAN1420QZ'

‘71!» En ﬂ89§§3 .’
'31:?“ 2007

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ma WWW“

ECOSYSTEM CONSEQUENCES AND SPATIAL DISTRIBUTION
OF SOIL MICROBIAL COMMUNITY STRUCTURE

By
Michel André Cavigelli

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Crop and Soil Sciences
1998

ABSTRACT

ECOSYSTEM CONSEQUENCES AND SPATIAL DISTRIBUTION
OF SOIL MICROBIAL COMMUNITY STRUCTURE

By
Michel André Cavigelli

The inﬂuence of biodiversity on ecosystem processes has been the subject of
considerable research effort and debate among plant and animal ecologists but the
ecosystem consequences of microbial diversity are largely unknown. I tested the
hypothesis that soil microbial diversity affects ecosystem function by evaluating the
effect of denitriﬁer community composition on nitrous oxide (N20) production. I used a
soil enzyme assay to evaluate the effect of oxygen concentration and pH on the activity of
denitriﬁcation enzymes responsible for the production and consumption of N20. By
controlling, or providing in non-limiting amounts, all known environmental regulators of
denitriﬁer N20 production and consumption, I created conditions in which the only
variable contributing to differences in denitriﬁcation rate and the relative rate of N20
production in soils from two ﬁelds in southwest Michigan was denitriﬁer community
composition. I found that the denitriﬁcation enzymes of the denitrifying communities
from the two ﬁelds differed substantially in their sensitivites to oxygen and pH,
indicating that the denitrifying communities in these two soils respond to environmental
regulators differently.

I also isolated denitrifying bacteria from these same soil samples and, for 31
representative isolates, measured the sensitivity of their nos enzymes, which catalyze the
reduction of N20 to N2, to low levels of oxygen. Cluster analysis of the cellular fatty acid
proﬁles of 93 denitrifying bacteria isolated from the agricultural ﬁeld and 63 from the
successional ﬁeld showed 27 denitrifying taxa with only 12 common to both soils. In

addition, I found substantial diversity in the degree of sensitivity of the isolates’ nos

enzymes to oxygen, indicating that the taxonomic diversity present among denitriﬁers in
these two soils is functionally signiﬁcant. These results demonstrate a clear potential for
differences in denitriﬁer community composition to affect differences in N20 production
among ecosystems, independent of direct environmental controls.

I also investigated the spatial distribution of microbial community structure along
a transect in a conventionally-tilled agricultural ﬁeld. I used fatty acid methyl ester
(FAME) proﬁle analysis, a rapid technique that allows processing of the large number of
samples required for spatial analyses of microbial conununity structure. I applied
principal components analysis to these data and showed that, while a majority of 167 soil
samples had similar FAME proﬁles, about 20 percent of samples had relatively low and
about 10 percent had relatively high bacterialzfungal ratios. Using semivariancc analysis I
was able to capture and describe small-scale patterns of microbial population
distributions. Where autocorrelation occured, it was generally at scales < 0.2 m, a scale

analagous to individual soil peds and rhizospheres.

To my parents, Agnes and Georges Cavigelli, who inspired critical intellectual inquiry,
attention to detail, and perseverance, and to Martha Tomecek who saw to it that I
completed this task with my sanity intact, and my spirit uplifted.

ACKNOWLEDGEMENTS

I would like to thank Phil Robertson, my advisor, for his support in this endeavor
and for his shared enthusiasm for and insights into science in general. Also, I wish to
thank my advisory committee members, Kay Gross, Mike Klug and Jim Tiedje for their
support. For patiently explaining many aspects of microbiological research to a soil
scientist/ecosystem ecologist, I thank Mary Ann Bruns, Joanne Chee-Sanford, Hal
Collins, Helen Garchow, Mike Kaufrnann, Sandy Marsh, Klaus Nuesslein, Rob Sanford,
and Pete Stahl. Per Ambus, Rob Sanford, and Jane Schuette gave helpﬁrl suggestions or
asked critical questions about my research, thus helping me reﬁne my methods. Wendy
Goodfn'end, Alan Tessier and Tom Willson helped me think through some of the
statistics. For other helpful discussions and for reviewing earlier versions of portions of
this dissertation, I thank Jane Boles, Tim Bergsma, Wendy Goodﬁiend, Heather
Reynolds, and John Rozum. For help in the lab with wieghing soil, making media,
managing data and maintaining GCs, I thank Amy Miller, Nelson Graves, Kerri Gorentz,
Helen Garchow, Sandy Marsh, Joe Picciano, Charity Wright, Shannan Gibb and Martha
Tomecek. Soil samples for the fourth chapter were jointly collected by Chen Ching
Chou, Deane Lehman, Lori Merrill, and myself. Mike Klug graciously provided space
and equipment necessary for portions of this project. John Gorentz helped with computer
questions and Carolyn Hammarskjold supplied many references from the campus library.

This research was supported ﬁnancially by the National Science Foundation
through a Doctoral Dissertation Improvement grant (DEB 9311380), the Center for
Microbial Ecology (DEB 9120006), and the Kellogg Biological Station Long-Term

Ecological Research Project in Row CrOp Agriculture (LTER; BSR 8702332), and by the
Michigan Agricultural Experiment Station.

I would like to extend a special thank you to three unique Michigan State
University programs for their support and inspiration: the Center for Microbial Ecology
and Jim Tiedj e, its director, for supporting an interactive research environment among
microbiologists working at many different levels of investigation and for supporting an
active scientiﬁc life that afforded many opportunities to interact with scientists from
around the country and the world; the KBS LTER program for providing a similar level
of professional development via contacts with scientists on campus and around the
country; and KBS for providing a stimulating and well equipped work environment.

Most importantly, I thank my best friend and wife, Martha Tomecek, who helped
me through the challenges of becoming a husband and father during the process of
making a dissertation. None of this would have been possible without her shared
wisdom, humor, insights, love and patience. I also thank Anna and Noah for their
patience with a too-often absent father. Finally, I thank Noah, Anna, Katia and Martha
for vigilantly reminding me, sometimes unknowingly, sometimes indirectly and
sometimes against my will, that science, despite the time and energy it requires,

represents only one portion of a complete life.

vi

PREFACE

Chapter 4 in this dissertation is a photocopy of a publication appearing in Plant

and Soil, volume 170, pages 99-113 (1995).

vii

TABLE OF CONTENTS

Page
LIST OF TABLES ............................................................................................................... x
LIST OF FIGURES ......................................................................................................... xiv
CHAPTER 1
ECOSYSTEM CONSEQUENCES AND SPATIAL VARIABILITY OF SOIL
MICROBIAL COMMUNITY STRUCTURE ..................................................................... 1
Ecosystem consequences of soil microbial community structure............................1
Spatial distribution of soil microbial community structure ..................................... 3
Study site and soils .................................................................................................. 5
CHAPTER 2
THE FUNCTIONAL SIGNIFICANCE OF DENITRIFIER COMMUNITY
COMPOSITION IN A TERRESTRIAL ECOSYSTEM ..................................................... 8
Introduction .............................................................................................................. 8
Materials and Methods ........................................................................................... 12
Results .................................................................................................................... 20
Discussion .............................................................................................................. 34
Conclusions ............................................................................................................ 38
CHAPTER 3
THE ROLE OF BIOTIC DIVERSITY IN RATES OF NITROUS OXIDE
CONSUMPTION IN A TERRESTRIAL ECOSYSTEM ................................................. 39
Introduction ............................................................................................................ 39
Materials and Methods ........................................................................................... 43
Results .................................................................................................................... 51

viii

Discussion .............................................................................................................. 71

Conclusions ............................................................................................................ 74
CHAPTER 4
FATTY ACID METHYL ESTER (FAME) PROFILES AS MEASURES OF SOIL
MICROBIAL COMMUNITY STRUCTURE ................................................................... 75
Introduction ............................................................................................................ 75
Materials and Methods ........................................................................................... 76
Results and Discussion .......................................................................................... 78
Conclusions ............................................................................................................ 87
CHAPTER 5 .
SUMMARY, SYNTHESIS, AND FUTURE RESEARCH POTENTIAL ....................... 90
APPENDIX ....................................................................................................................... 95
LIST OF REFERENCES ................................................................................................. 132

LIST OF TABLES

Chapter 1

Table 1.1 - Soil properties for the conventionally-tilled agricultural ﬁeld and the never-
tilled successional ﬁeld at the KBS LTER site. ................................................. 7

Chapter 2

Table 2.1 - Denitriﬁcation rate at low oxygen concentration (equilibrium slurry
concentration, 0.27 pmol ' L") with increasing shaking speed for soil ﬁom one
ﬁeld replicate of the agricultural ﬁeld at the KBS LTER site. ........................ 17

Table 2.2 - Nitrous oxide production rates with and without acetylene under anaerobic
and microacrobic conditions, and the ratio between anaerobic and
microacrobic rates for soils from the conventionally-tilled agricultural ﬁeld
and the never-tilled successional ﬁeld at the KBS LTER site. ........................ 21

Table 2.3 - Total denitriﬁcation for soil from the agricultural ﬁeld at the KBS LTER site
with increasing rate of nitrate addition. ........................................................... 25

Table 2.4 - Calculated slurry nitrate concentrations at two pH levels at the point of initial
nos activity during preincubations for soils from the conventionally-tilled
agricultural ﬁeld and the never-tilled successional ﬁeld at the KBS LTER site
.......................................................................................................................... 27

Table 2.5 - F values of three-way AN OVAs to determine effects of site, pH, and oxygen
on denitriﬁcation rate; net N20 production rate; and erO, the relative rate of
N20 production for soils from the conventionally-tilled agricultural ﬁeld and
the never-tilled successional ﬁeld at the KBS LTER site. ............................... 3O

Chapter 3
Table 3.1. Reference strains used to compare cellular fatty acid proﬁles against those of
denitrifying bacteria isolated from the conventionally-tilled agricultural ﬁeld
and the never-tilled successional ﬁeld at the KBS LTER site ......................... 46
Table 3.2. Mean number of anaerobic heterotrophic bacteria viable on R2A* agar, mean
number of gas-producers in R2A* broth, and mean number of conﬁrmed
denitriﬁers for soils from the conventionally-tilled agricultural ﬁeld and the
never-tilled successional ﬁeld at the KBS LTER site ........................... - ........... 52
Table 3.3. Denitrifying bacteria isolated from the conventionally-tilled agricultural ﬁeld
and the never-tilled successional ﬁeld at the KBS LTER site, and reference“
strains, grouped by taxa deﬁned by cluster analysis (Figure 3.1) .................... 55
Table 3.4. The effect of oxygen on N20 consumption rates for 31 denitrifying isolates
representing 20 different taxa isolated from the conventionally-tilled
agricultural ﬁeld and the never-tilled successional ﬁeld at the KBS LTER site.
.......................................................................................................................... 63
Table 3.5. Exponential decay constants, k, used to measure the sensitivity of nos to
oxygen for seven isolates measured at two different points in time aﬁer N20
consumption began. These experiments test whether variables that are likely
to covary with time of sampling affect the inﬂuence of oxygen on N20
consumption rate. Results show that sampling time did not inﬂuence k
values ............................................................................................................... 64
Table 3.6. Exponential decay constants, k, used to measure the sensitivity of nos to
oxygen for isolates belonging to taxa with two test isolates. Results show that
there were no differences in the k values between isolates belonging to the

same taxa .......................................................................................................... 65

xi

Table 3.7. Exponential decay constants, k, for the seven isolates belonging to taxon 30

Table 3.8.

and t test results to compare k values among the seven isolates. Results show
that there were no differences among k values for the seven isolates belonging
to taxon 3O ........................................................................................................ 67
ANCOVA table to determine the effect of isolate on the sensitivity of nos

activity to oxygen, k (main effect, isolate; covariate, oxygen) ........................ 68

Chapter 4 (Plant and Soil 170199-113)

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Marker fatty acids. From Erwin (1973), White (1983), Harwood and Russell
(1984), Jantzen and Bryn (1985), and Vestal and White (1989) ..................... 77
Fatty acids found in soil and in plated communities. Those fatty acids found
in both communities are indicated in boldface. Fatty acids with retention
times > 17 mins on the GC column and removed for the whole soil analyses
are listed as a separate group ............................................................................ 80
Proportion of variance accounted for by eigenvalues of the correlation
matrices with values 3 l for principal component analyses ............................ 81
FAME proﬁle characteristics most inﬂuential in distinguishing the soil
samples represented by stars from those represented by open circles in Figures
2A (whole soil communities) and 5A (plated communities). Those FAMEs
that distinguished soil samples from both soil and plated communities are
indicated in bold. All FAMEs which had loadings of more than |0.20| (an
arbitrary threshold) in the relevant eigenvectors are included in this table. The
value of the loadings are indicated in columns following each FAME ........... 83
Parameters for variograms of FAMEs that exhibited spatial autocorrelation,
including variograms presented in Figure 4. C/(CO+C) is the proportion of

population variance due to spatial structure and A0 is the range ..................... 86

xii

Appendix

Table A1. FAME proﬁles for 35 reference strains and 156 denitrifying bacteria isolated
from soils from the conventionally-tilled agricultural ﬁeld and the never-tilled
successional ﬁeld at the KBS LTER site. ........................................................ 96

xiii

LIST OF FIGURES

Chapter 2

Figure 2.1 - Denitriﬁer community enzyme induction curves at two pH levels for soils
from the conventionally-tilled agricultural ﬁeld and the never-tilled
successional ﬁeld at the KBS LTER site. Nitrous oxide reductase activity is
indicated by greater N20 production in the presence than in the absence of
acetylene. Values are means and error bars are i 1 SE (n = 3 for agricultural
ﬁeld; 11 = 2 for successional ﬁeld). Note different ranges for x axes. ........... 23

Figure 2.2 - Denitriﬁcation (A and B) and N20 production rates (C and D) at four oxygen
concentrations and two pH levels for soils ﬁom the conventionally-tilled
agricultural ﬁeld and the never-tilled successional ﬁeld at the KBS LTER
site. Values are means and error bars are i 1 SE (n = 3 for the agricultural
ﬁeld; 11 = 2 for the successional ﬁeld). ........................................................... 29

Figure 2.3 - Relative rate of N20 accumulation, rNZO, at four oxygen and two pH levels
for soil from the conventionally-tilled agricultural ﬁeld and the never-tilled
successional ﬁeld at the KBS LTER site. Values are means and error bars are
i 1 SE (n = 3 for the agricultural ﬁeld; 11 = 2 for the successional ﬁeld). ...... 33

Chapter 3

Figure 3.1 - Dendrograrn, based on cellular fatty acid proﬁles, of 36 reference strains and
156 denitrifying bacteria isolated from soils from the conventionally-tilled
agricultural ﬁeld and the never-tilled successional ﬁeld at the KBS LTER

site. Taxonomic groupings at a coarse scale (0.53 Euclidean units) are

xiv

indicated by the letters A, B, and C. Taxonomic clusters at a ﬁner scale
(33 species-level taxa) were deﬁned by grouping isolates at the Euclidean
distance represented by the dashed line (0.13 Euclidean units). These
clusters are identiﬁed by the . numbers listed to the right. Isolates comprising
each taxa are listed in Table 3.3 .................................................................... 54
Figure 3.2 - A. Typical patterns of nitrous oxide production by denitrifying isolates
grown in batch culture in the presence (closed circles) and absence (open
symbols) of acetylene. Nitrous oxide reductase (nos) activity is suggested by
the combination of high N20 production in the presence of acetylene and no
N20 production in the absence of acetylene. Nos activity is conﬁrmed by ‘
rapid N20 consumption following injection of 70 mo] N20°L'l in the
absence of acetylene (open squares). B-D. Atypical N20 production patterns
in the presence of acetylene for isolate numbers 46 (B), 77 and 85 (C), and 89
(D) .................................................................................................................. 59
Figure 3.3 - Typical nitrous oxide consumption pattern for denitrifying isolates in
incubation vials to which 5 mL of 1.01 percent N20 was added. There was
no net N20 production prior to rapid consumption ....................................... 62
Figure 3.4 - Sensitivity of nos to oxygen for 31 denitrifying isolates representing 20 taxa
isolated from the conventionally-tilled agricultural ﬁeld and the never-tilled
successional ﬁeld at the KBS LTER site. The sensitivity of nos to oxygen
was quantiﬁed by plotting the natural log of N20 consumption rate against
oxygen level. These slopes are equivalent to the exponential decay constant,
k, for the untransformed data that are presented in Table 3.4. Error bars are i
1 SE. Taxa numbers below the x-axis are the same as in Figure 3.1 and
Tables 3.3 and 3.4. Numbers in parentheses above the x-axis are the isolate
number for those taxa with more than one test isolate. R* refers to the

XV

reference strain P. ﬂuorescence F ATCC 17513. Isolate number is not

included for those taxa containing only one test isolate ................................ 70

Chapter 4 (Plant and Soil 170299-113)

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Means and standard deviations (represented by error bars) of FAMEs for ﬁve

. subsamples from one homogenized sample of soil taken from a corn ﬁeld ..79

PCA plot and biplots for soil community FAME proﬁles. A. PCA plot for
whole soil community FAME proﬁles containing 46 FAMEs. B. Biplot for
whole soil community FAME proﬁles containing 14 F AMEs selected using
PCA. C. Biplot for whole community FAME proﬁles containing 10 groups
ofFAMEs. Parameters are listed in Table 382
Dendrograrns derived from cluster analyses for whole soil communities. A.
Euclidean clustering. UPGMA linkage. B. Average taxonomic distance
clustering, single linkage. The same symbols used in Figures 2 and 5 are
used for each soil sample ............................................................................... 84
Selected variograms for soil community F AMEs analyzed with a minimum
step of 0.07 m (A) and 0.03 to 0.05 m (B), and plated commrmity FAMEs
analyzed with a minimum step of 0.07 m (C) and 0.03 to 0.05 m (D). The x
axes are distance (m) and the y axes are semivariancc (y) ............................ 85
PCA plot and biplots for plated community FAME proﬁles. A. PCA plot for
plated community FAME proﬁles containing 41 FAMEs. B. Biplot for plated
community FAME proﬁles containing 11 FAMEs selected using PCA.

Parameters are listed in Table 3 ..................................................................... 87

CHAPTER 1
ECOSYSTEM CONSEQUENCES AND SPATIAL VARIABILITY OF SOIL
MICROBIAL COMMUNITY STRUCTURE

In this dissertation I address, from a microbial perspective, two important
ecological issues that have received considerable attention among plant and animal
ecologists: the role of biodiversity on ecosystem ﬁmction (e. g. Mooney et a1. 1995,
Hooper and Vitousek 1997, Huston 1997, Grime 1997, Chapin et a1. 1998) and the spatial
variability of populations and communities (e.g. Dunning et a1. 1992, Kadmon 1993,
Hunter et a1. 1992, Turner 1989)). These questions have received less attention from
microbial ecologists due, in part, to a general assumption that microbial diversity is
ﬁmctionally redundant (e. g. Meyer 1993, Beare et al. 1995, Heal et al. 1996, but see
Schulze and Mooney 1993) and to methodological limitations in studying soil

microorganisms.

Ecosystem consequences of soil microbial community structure

The prevailing belief that microbial diversity is ﬁmctionally redundant is based on
at least two points. One point is that there is an enormous diversity of microorganisms
responsible for many ecosystem ﬁrnctions, thus suggesting that there must be signiﬁcant
redundancy within functional groups (c.g. Beare et a1. 1995). The second point is that
ecosystem-level process models are able to predict some microbially-mediated ecosystem
processes without considering the composition of the microbial community (6. g. Conrad
1996). Mathematical models of soil C02 ﬂux are commonly cited to support this view

(e. g. Schimel 1995). On the other hand, ecosystem-level production and consumption of
1

other trace gas ﬂuxes, including that of N20, have been harder to predict using solely
environmental data (e. g. McGill et a1. 1981; McConnaughey and Bouldin 1985; Rolston
et al. 1984; Johnsson et a1. 1987; Arah and Smith 1989; Parton et a1. 1988; Li et al.
1992a, b; Ojima et a1. 1992; Smith et a1. 1993; Parton et a1. 1995). Also, there is growing
evidence from the microbiological literature that different organisms produce
endproducts at different rates even when grown under identical situations (e. g. Korner
1993, Robertson et a1. 1995, Conrad 1996, Ka et a1. 1997). Such ﬁndings suggest that the
regulation of ecosystem processes that are mediated by microbes may in fact be‘affected
by -- and reﬂect -- the community composition of ﬁrnctional groups. Some
microbiologists and ecosystem ecologists are beginning to explore this possibility. For
example, Conrad (1996) conluded an exhaustive review on the potential effects of
microbial diversity on trace gas production (H2, CO, CH4, OCS, N20, and NO) with the

comment,

...it remains unclear whether trace gas ﬂux at the ecosystem level is ﬁnally
controlled by a few processes only or whether changes in the diversity and
composition of microbial species are also important. However, by
analogy to ecological studies of plant and animal communities...I would
dare to predict that microbial species diversity is an important factor in
ecosystem function.

Schimel (1995) likewise postulated that for such specialized processes as denitriﬁcation,
nitriﬁcation, and methane oxidation, microbial diversity is likely to affect ecosystem
function because of the relatively limited number of organisms capable of carrying out
these transformations, i.e. because redundancy is lower among these specialized groups.

To address the question of ecosystem consequences of soil microbial community
structure, I studied nitrous oxide (N 20) production by soil denitrifying bacteria. Nitrous
oxide is a greenhouse gas and natural catalyst of stratospheric ozone degradation

(Rasmussen and Khalil 1986, Cicerone 1987), and its concentration in the atmosphere is

increasing annually at about 0.8 parts per billion by volume (Houghton et a1. 1996).
Denitriﬁers are oﬁen the dominant source of N20 from soils (Firestone and Davidson
1989) and they are the only organisms that can consume N20 (e.g. Conrad 1996).

I used two approaches to study the role of soil denitriﬁer community composition
on N20 production and consumption. For the research presented in Chapter 2, I modiﬁed
the standard denitriﬁcation enzyme assay (Smith et a1. 1978, Smith and Tiedje 1979,
Tiedje 1994). This assay is designed to measure the activity of denitriﬁcation enzymes
by providing unlimited concentrations of denitriﬁcation substrates (carbon and nitrate)
under denitriﬁcation (anaerobic) conditions at constant temperature and moisture. I
modiﬁed this assay by controlling all additional environmental factors that regulate the
relative rate of N20 accumulation (pH, oxygen, denitriﬁcation enzyme status), thus
creating conditions wherein any observed differences in the relative rate of N20
accumulation (erO) can only be due to differences in denitriﬁer community
composition. In order to relate ﬁndings from these soil enzyme assays to the physiology
of the organisms responsible for N20 production and consumption, I also isolated
denitrifying bacteria from these same soil samples and measured the sensitivity to low
oxygen concentrations of their nos enzymes, which reduce N20 to N2 during
denitriﬁcation. I also described the denitriﬁer community composition of both ﬁelds
based on fatty acid methyl ester (FAME) proﬁle analysis of 156 isolates, together with 36

reference strains. These results are presented in Chapter 3.

Spatial distribution of soil microbial community structure
The spatial distribution of soil microbial community structure has been described,
usually qualitatively, at a number of different scales: among aggregates of different sizes

(e.g. Hattori 1988), in the rhizosphere compared to bulk soil (e.g. Lynch 1986, Foster

1988), with soil depth (e.g. Weier and MacRae 1992), and in different size soil pores (e.g.
Foster 1988). The ecological relevance of most studies on the spatial variability of soil
microbial community structure, however, is limited by at least two factors: 1) many are
based on isolating soil microorganisms, a technique that samples less than 10 percent of
the total soil populations (e.g.. Torsvik et al. 1990a,b) and not necessarily the
numerically-dominant ones (e. g. Gorlach 1994); and 2) since conventional techniques
(isolating microorganisms or electron microscopy) are time-consuming, descriptions are
usually based on a relatively small number of samples. Thus, the scale of variability has
rarely been quantiﬁed and the proportion of total variation attributable to spatial factors
cannot be determined. To link spatial patterns of soil microbial community structure to
other relevant ecological parameters such as plant population and community patterns
requires spatially explicit statistical tools, such as geostatistical analysis (Robertson and
Gross 1994).

To address both these issues I used a novel biochemical technique, fatty acid
methyl ester (FAME) analysis -- a rapid technique that does not require the culturing of
microorganisms -- to characterize soil microbial community structure, and multivariate
and geostatistical analyses to interpret FAME proﬁle patterns. Soil samples were taken
along a transect in a conventionally-tilled corn ﬁeld in a nested design to capture both
small and large scale spatial strucutre.

FAME is a relatively rapid technique that takes advantage of differences in the
cellular fatty acid content in different microbial taxa. Fatty acids are ﬁrst extracted from
environmental samples, methylated and analyzed as FAMEs on a gas chromatograph
(MIDI 1992). The resulting FAME proﬁles can be interpreted using knowledge of
speciﬁc fatty acids that are representative of particular microbial taxa (Vestal and White

1989). Originally developed for rapid, inexpensive identiﬁcation of medical isolates

(MIDI 1992), the method was adapted to environmental samples by DC. White and
colleagues in Tennessee. Their investigations, however, were almost exclusively limited
to sediments (e.g. Bobbie and White 1980, Findlay and White 1983, Vestal and White
1989). A research group working at Michigan State University’s WK. Kellogg
Biological Station (KBS) was one of the ﬁrst to apply FAME analysis to soil samples and
to questions of soil ecology. The research presented in Chapter 4 is the ﬁrst to apply
geostatistical and multivariate analyses to address ecological questions using FAME

analysis of environmental samples.

Study site and soils

All three reports presented in this dissertation were conducted on soils sampled
from two ﬁelds at the Long-Term Ecological Research (LTER) site in agricultural
ecology at KBS, which is located in southwest Michigan in the northern part of the US.
corn belt. The area receives about 860 m y'1 of precipitation, evenly distributed
throughout the year. Mean annual temperature is 9.4 °C. Solar and sky radiation are low,
with appreciable precipitation occurring on about 100 days y'l. Soils developed on the
pitted outwash plain following the Wisconsin glaciation 10,000 years ago. Soils at this
study site are Typic Hapludalfs (ﬁne-loamy, mixed, mesic) of moderate fertility that
developed under forest vegetation.

For research presented in Chapters 2 and 3 I collected soil samples in the fall of
1994 from a conventionally-tilled agricultural ﬁeld and a never-tilled successional ﬁeld.
The agricultural ﬁeld had been in a com-soybean-wheat rotation since 1988 and in
various ﬁeld crop rotations for the prior century, managed in accordance with regional
agronomic practices. The successional ﬁeld was cleared of trees in 1960 and has since

been mowed annually or biannually. Dominant plant species in the successional ﬁeld at

the time of sampling, expressed as percent of total live biomass, were Arrhenatherum
elatius (L.) Beauv.ex J .-BC Presl (tall oatgrass, 15 percent), Elaeagnus umbellata Thunb.
(autumn olive, 15 percent), Solidago canadensis L. (common goldenrod, 9 percent),
Rubus allegheniensis T.C. Porter (blackberry, 9 percent), Monarda ﬁstulosa L. (beebalm,
8 percent), Bromus inermis Leyss. (smooth brome, 7 percent), Poa pratensis L.
(Kentucky bluegrass, 7 percent), P. compressa L. (Canada bluegrass, 6 percent), and
Dactylis glomerata L. (orchardgrass, 6 percent). Management of these sites is more fully
described on the world wide web (http://kbs.msu.edu/lter/).

The two ﬁelds have geomorphically similar soils that differ in pH, bulk density,
total carbon, total nitrogen, and inorganic nitrogen pools (Table 1.1) -- factors that I
judged likely to inﬂuence denitriﬁer community composition. From each of three ﬁeld
replicates of the two sites, I composited thirty 2.5 cm soil cores (10 cm depth) for a total
of six soil samples. Samples were stored on ice until returned to the lab, at which point
they were homogenized through 4 mm sieves and stored at 4°C until subsarnpled for
enzyme assays and isolate work.

The research presented in Chapter 4 was conducted on soil samples collected in
the summer of 1991 ﬁ'om the same conventionally-tilled agricultural ﬁeld described
above. Sampling methods are described in that Chapter, which is a reprint of Cavigelli et

a1. (1995).

Table 1.1. Soil properties for the conventionally-tilled agricultural ﬁeld and the never-

tilled successional ﬁeld at the KBS LTER site.

 

 

Soil Property Agriculturziﬁeld Successional ﬁeld t testa—
Texture sandy loam sandy loam

pH" 6.59 i 0.04 5.70 2: 0.11 **
Moistureb (g - g soil") 10.49 : 0.65 11.24 i 0.56 ns

Bulk densityc (g - cm'3) 1.65 i 0.05 1.35 i 0.02 *

Total carbond (g c- 100 g soil") 0.77 i 0.07 1.97 i 0.12 **

Total nitrogend (g N ° 100 g soil") 0.077 : 0.006 0.166 i 0.009 **
Nitrate-N‘ (pg N - g soil") 4.57 i 0.95 0.93 i 0.37 *
Ammonium-Ne (pg N - g soil“) 2.73 i 0.76 8.75 i 0.35 **

 

Note: Values are means i 1 SE for three ﬁeld replicates; signiﬁcance values are based on
t tests. All data except pH and moisture are from the KBS LTER website
(http://kbs.msu.edu/lter/). The soils of this site are described in greater detail by Collins
et a1. (1998).

3 us, not signiﬁcant, P > 0.1; * P < 0.05; ** P < 0.01.

b sampled 21 September 1994 to a depth of 10 cm.

c sampled prior to plowing, spring 1996 to a depth of 15 cm.

d sampled 29 August 1994 to a depth of 25 cm.

c sampled 21 September 1994 to a depth of 15 cm.

Chapter 2
THE FUNCTIONAL SIGNIFICANCE OF DENITRIFIER COMMUNITY
COMPOSITION IN A TERRESTRIAL ECOSYSTEM

INTRODUCTION

The signiﬁcance of biodiversity to ecosystem function has emerged as a major
ecological issue in recent years, particularly among ecologists working with
macroorganisms (e. g. Mooney et al. 1995, Grime 1997, Hooper and Vitousek 1997,
Huston 1997, Chapin et al. 1998). At question is the degree to which species diversity or
individual species affect overall ecosystem function and the delivery of ecosystem
services. This question is equally relevant for microbial taxa (e. g. Schimel 1995, Conrad
1996), especially in light of the tremendous apparent diversity among soil
microorganisms. For example, recent estimates of microbial species richness, based on
novel molecular techniques, suggest that there can be as many as 4-10 x 103 different
bacterial taxa in a single soil sample (Torsvik et al. 1990a, Klug and Tiedje 1994).

Within broad taxonomic bounds, microbial diversity clearly matters a great deal:
only certain taxa appear to carry out particular biogeochemical processes. It is unknown,
however, whether diversity within the broad ﬁmctional groups that perform these
processes -- groups such as nitriﬁers, denitriﬁers, and methanotrophs -- is important.
While symbiotic microorganisms can have high host-speciﬁcity (e. g. Alexander 1985,
Allen et a1. 1995), diversity within most microbial functional groups is often assumed to
be functionally redundant (e.g. Meyer 1993, Beare et al. 1995, Heal et a1. 1996). In fact,

in both schematic and mechanistic models of nutrient cycling, functional groups are

depicted and treated as black boxes that transform inputs to outputs at rates determined
solely on a speciﬁc set of rate-limiting environmental factors.

Enzyrnological studies of isolated microorganisms, on the other hand, often reveal
great variability in enzyme immunochemical cross-reactivity, efﬁciency, and regulation
within functional groups (e.g. Komer 1993, Robertson et al. 1995, Conrad 1996, Ka et al.
1997, this volume Chapter 2). Such results suggest that the regulation of ecosystem
processes that are mediated by microbes may in fact be affected by -- and reﬂect -- the
community composition of functional groups.

Denitriﬁers provide an excellent microbial model for studying questions related to
the functional signiﬁcance of biotic diversity: they are among the most diverse groups of
bacteria in terrestrial ecosystems and -- because of their importance to nitrogen (N)
cycling ~— among the best studied (Tiedje 1988, Zumft 1992). Denitriﬁcation, which is
carried out solely by denitrifying bacteria, is the conversion of nitrate (N03) and nitrite
(N02) to the nitrogen gases nitric oxide (NO), nitrous oxide (N20), and dinitrogen (N 2).
Thus, denitriﬁcation can have a direct impact on soil nitrogen availability and, ultimately,
on net primary production in many terrestrial and coastal margin ecosystems.
Denitriﬁcation is also a major source of atmospheric N20 (Firestone and Davidson 1989),
an important greenhouse gas and a natural catalyst of stratospheric ozone decay
(Rasmussen and Khalil 1986, Cicerone 1987). The atmospheric concentration of N20 is
increasing annually at about 0.8 parts per billion by volume (Houghton et al. 1996) and
the global N20 budget is far from balanced (Davidson 1991, Robertson 1993).
Denitriﬁcation’s role in this increase is unclear.

Denitriﬁers are facultative anaerobes, capable of shunting electrons from electron
transport phosphorylation (ETP) to nitrogen oxides when oxygen becomes limiting
(Tiedje 1988). The four enzymes that link ETP to nitrogen oxide reduction (nitrate

1O

reductase, nar; nitrite reductase, nir; nitric oxide reductase, nor; and nitrous oxide
reductase, nos) are usually induced sequentially under anaerobic conditions (Tiedje
1994»

N03' —> N02’ —> NO —> N20 —> N2.
nar nir nor nos

Since N20 is produced by nor and consumed by nos, N20 accumulates during
denitriﬁcation under two sets of conditions: 1) aﬁer nor but before nos is induced
(Firestone and Tiedje 1979, Dendooven and Anderson 1994), and 2) following induction
of the entire denitriﬁcation pathway when environmental conditions inhibit nos activity to
a greater extent than they inhibit nor activity (Betlach and Tiedje 1981). Temperature,
oxygen, pH, available organic carbon, and nitrate and/or nitrite inﬂuence N20 production
and consumption rates via their inﬂuence on both denitriﬁcation enzyme induction and
activity (Tiedje 1988, Firestone and Davidson 1989).

The rate of denitn'ﬁer N20 accumulation is thus a product of both denitriﬁcation

rate [A(NZO+N2)] and the relative rate of N20 production,

AN20
A(NzOq’Nz)

or, for simplicity, erO. Within the small range of low oxygen concentrations under
which denitriﬁcation usually occurs, denitriﬁcation rate generally decreases with
increasing oxygen and decreasing pH. rNZO, on the other hand, generally increases with
increasing oxygen, decreasing pH, and decreasing carbonznitrate ratio (e. g. Firestone et
al. 1979, 1980; Firestone and Davidson 1989).

The major environmental controls on deniuiﬁcation rate and rNZO have been
incorporated into mechanistic models, but these models are, in general, poor predictors of
in situ N20 ﬂux rates (e.g. McGill et a1. 1981; McConnaughey and Bouldin 1985;
Rolston et al. 1984; Johnsson et a1. 1987; Arah and Smith 1989; Parton et al. 1988; Li et
al. 1992a, b; Ojima et a1. 1992; Smith et al. 1993; Parton et al. 1995). Modeling

11

difﬁculties have been attributed to the high degree of spatial and temporal variability of
environmental conditions controlling denitriﬁcation and N20 emissions (6. g. Arah and
Smith 1990). An additional, untested explanation is that the soil environment harbors
denitriﬁer populations with denitriﬁcation enzymes, especially nor and nos, that differ in
their sensitivity to environmental variables. If this is the case, our ability to predict
changes in local and global N20 ﬂuxes may hinge on an understanding of denitriﬁer
community composition and population dynamics.

I show in Chapter 2 that the composition of the denitrifying community in
contrasting soils from an agricultural ﬁeld and a successional ﬁeld differ and that there is
great diversity in the sensitivity of nos to oxygen among different denitriﬁers cultured
from these soils. I test here the hypothesis that denitriﬁer community composition can
affect denitriﬁcation rate and rNZO using a soil enzyme assay designed to evaluate the
effect of oxygen concentration and pH on the activity of all denitriﬁcation enzymes in the

entire denitriﬁer conununity, including those populations that may not be culturable.

12

MATERIALS AND METHODS

My overall strategy is to compare the sensitivity of denitriﬁcation enzymes to
very low oxygen concentrations in soils from two sites after controlling for all other
environmental factors that regulate the denitriﬁcation rate and rNZO. These factors
include temperature, pH, carbonznitrate ratio, substrate diffusion, and denitriﬁcation
enzyme induction status. I am aware of no other important environmental factors that can
affect denitriﬁcation rate and rNZO in soils. Any observed differences in denitriﬁcation
rate and rNZO between soils under these controlled conditions, then, is evidence that soil
denitrifying commtmities respond differently to the same environmental conditions, i.e.

that denitriﬁer community composition inﬂuences N20 production.

Study site and soils
I collected soil samples in the fall of 1994 from a conventionally-tilled
agricultural ﬁeld and a never-tilled successional ﬁeld at the Long-Term Ecological
Research (LTER) site in agricultural ecology at KBS. The study site and soils are
described in Chapter 1, including Table 1.1. From each of three ﬁeld replicates of the two
sites, I composited thirty 2.5 cm soil cores (10 cm depth) for a total of six soil samples.
Samples were stored on ice until returned to the lab, at which point they were

homogenized through 4 mm sieves and stored at 4°C until subsampled for enzyme assays.

Preincubation: Denitriﬁer community enzyme induction
Prior to measuring denitriﬁer community enzyme sensitivity to very low oxygen
concentrations, I induced denitriﬁcation enzymes using an anaerobic slurry preincubation
and conﬁrmed the activity of nos, the last enzyme in the pathway. This preincubation

was necessary to ensure that enzyme activity measurements are not confounded by the

13

enzyme induction status of the community at the time of sampling (6. g. Smith and Tiedje
1979, Dendooven and Anderson 1994).

I measured 3.0 g of each soil into two sets of 38 mL serum vials, sealed the vials
using butyl rubber septa and crimp seals, and removed oxygen by ﬂushing for about three
min with ultra high purity (UHP; 99.995%) N2 gas. I added 3.0 mL of degassed water by
syringe and equilibrated the headspace pressure to atmospheric pressure using a needle
attached to a syringe barrel containing a small amount of water. To one set of vials, I
added 3.6 mL of acetylene, which blocks the reduction of N20 to N2 by nos (Balderston
et al. 1976, Yoshinari and Knowles 1976). I made oxygen-free acetylene by adding
degassed water to a sealed serum vial containing calcium carbide in an evacuated N2
atmosphere. After adding acetylene to vials I re-equilibrated the headspace pressure and
then monitored headspace gases during the following 30-80 hours by sampling with a
gas-tight needle and syringe (0.1 mL sample volume using a 1 mL syringe). Comparing
N20 accumulation in the two sets of vials (with and without acetylene) allowed me to
monitor nos activity (Smith et al. 1978).

I measured both N20 and 02 using an HP 589011 gas chromatograph (GC;
Hewlett Packard, Rolling Meadows, IL) equipped with dual Poropak Q columns and an
electron capture detector (ECD). Gas concentrations in the headspace were corrected for
dissolution in the slurry solution using Bunsen coefﬁcients (Tiedje 1994). I used separate
syringes (and standard curves) for those vials that received acetylene to avoid acetylene
cross-contamination.

To assess whether nos induction regulation by nitrate or nitrite was different in
these two communities, I calculated the nitrate concentration in the two soils at the
approximate time of nos induction. I ﬁrst calculated the amount of nitrate present in each

soil at the beginning of the induction assays by converting the amount of N20 that

14

accumulated in the presence of acetylene to the amount of NO3'-N required to produce
that amount of N20 (Lensi et al. 1985, Christensen and Tiedje 1988, Binnerup and
Sorensen 1992, Hojberg et al. 1994). I then subtracted from initial NO3'-N concentrations
the amount of NO3'-N converted to N20 at the time of nos induction. These calculations
assume that N20 accumulation in the presence of acetylene is nitrate-limited. To test this
assumption, I added slurry solution containing 17, 33, 50, and 67 pg NO3'-N ' g soil '1 to a
separate set of soil subsamples. After denitriﬁcation had peaked in vials to which 67 pg
NO3'-N ' g soil '1 had been added, I added an additional dose of 67 pg NO3'-N ' g soil '1. I

monitored denitriﬁcation in all vials every 2 to 10 h for 100 h.

Denitrifier community enzyme activity assay

I measured the activity of nos under very low oxygen concentrations by
conducting short-term slurry incubations in which all environmental factors controlling
denitriﬁcation rate and erO were controlled or provided in non-limiting amounts. This
method is a modiﬁcation of the denitriﬁcation enzyme assay (Smith et al. 1978, Smith
and Tiedje 1979, Tiedje 1994).

After conﬁrming nos activity under preincubation conditions, I ﬂushed
preincubation vials that had not been exposed to acetylene with UHP N2 while shaking
until N20 in the headspace was below the GC detection limit. I randomly assigned each
vial to one of two sets. To one set I added 3.6 mL acetylene (about 10 kPa headspace
concentration) and equilibrated headspace pressure to atmospheric pressure. I then added
0, 0.05, 0.10 or 0.15 mL of 100 percent oxygen to all vials together with 0.1 mL of a
degassed solution of succinate and nitrate, to a ﬁnal concentration of 1 mM each. These

levels are generally considered non-limiting for soil denitrifying communities (Tiedj e

15

1994). Vials were shaken at 400 rpm using an orbit shaker (Model 3520, Lab-Line
Instruments, Melrose Park, IL) immediately after addition of the degassed solution.

I took vial headspace samples every 7-10 min after the start of the incubation and
measured N20 using the GC. Rates of net N20 production (ANZO; measured in the
absence of acetylene) and denitriﬁcation [A(NZO+N2); measured in the presence of 10
percent acetylene] were calculated ﬁom linear regressions of at least four points sampled
within the ﬁrst 60 min of incubation. In all assays, regression coefﬁcients were greater
than 0.90, indicating that there was no de novo enzyme synthesis during these short-term
incubations. I did not use chloramphenicol, an inhibitor of protein synthesis that is
sometimes necessary for similar incubations (Tiedje 1994).

Denitriﬁcation rates measured at the four oxygen levels provide a measure of the
sensitivity to oxygen of the enzymes leading to the production of N20 (nar, nir and nor).
Net N20 production rates measured at the four oxygen levels provide an integrated
measure of both N20 production and consumption at low oxygen concentrations. The
sensitivity of N20 consumption, or nos activity, to oxygen can therefore be measured as
erO, the ratio of net N20 production to denitriﬁcation rate [ANzO/A(NZO+N2)] at
different oxygen levels.

I maintained oxygen concentrations in the slurry solutions at 0.17, 0.27, and 0.45
pmol ' L'I for the 0.05, 0.10, and 0.15 mL oxygen additions, respectively. These levels are
well below the estimated 10 umol ' L'l threshold for denitriﬁcation activity (Tiedje 1988).
I found no evidence for signiﬁcant oxygen consumption during these short term
incubations, i.e. oxygen concentrations, corrected for GC ﬂuctuations, did not vary during
incubations. For convenience I report oxygen concentrations as mL of oxygen (100

percent) added per vial.

16

I had earlier tested the effect of shaking speed on denitriﬁcation rate in similar
vials to which 0.3 mL oxygen had been added. Results from this test showed that
denitriﬁcation rate decreased with increasing shaking speed up to 300 rpm (Table 2.1),
suggesting that oxygen difﬁrsion from gas to water phase was limited at low shaking
speeds. Shaking vials at 400 rpm alleviated this problem, i.e. oxygen concentrations in
the soil slurry solution were in equilibrium with the gas phase at sufﬁciently high shaking
speeds.

To test whether vigorous shaking affected bacterial viability, I compared bacterial
plate counts from unamended aerobic slurries of one ﬁeld replicate from the agricultural
ﬁeld after 4 h of vigorous (400 rpm on orbit shaker) or gentle (using a hematology ’
chemistry mixer, Model 346, Fisher Scientiﬁc) shaking. I followed standard plate count
techniques (Zuberer 1994) using a modiﬁed R2A medium (R2A*, Chapter 2) and
conducted the experiment using three replicate vials. I counted all colonies on plates with
30 to 300 colonies after 24 h of growth at room temperature. Mean plate counts were

log-transformed and compared using the Student’s t test (Ott 1984).

pH of enzyme induction and activity assays
I conducted all enzyme induction and activity assays at both native pH and at a
pH level adjusted to that of the other soil. These treatments were necessary to control for
effects of pH on denitriﬁcation rate and rN20 (e. g. Firestone et a1. 1980). I used
phosphate buffer (pH 5.5) to decrease the pH of the soil from the agricultural ﬁeld to 5.74
i 0.03 (mean 1; SE) and calcium carbonate (0.033 g ' g soil'l) to increase the pH of the soil
from the successional ﬁeld to 6.89 i 0.12 (mean 3: SE). To test whether there were any

cherrrical effects of phosphate buffer or calcium carbonate on enzyme induction or

17

Table 2.1. Denitriﬁcation rate at low oxygen concentration (equilibrium slurry
concentration, 0.27 umol ° L!) with increasing shaking speed for soil from one ﬁeld

replicate of the agricultural ﬁeld at the KBS LTER site.

 

Denitriﬁcation rate

 

 

 

Shameed (n2 N2O= soirimi)
0 1.93 : 0.03a
200 1.19 : 0.02b
300 0.96 : 0.03c
350 0.95 i 0.046
400 0.99 i 0.050

 

Note: Values are means i 1 SE (n = 3). Above 300 rpm the rate of oxygen diffusion ﬁ'om
gas to water phase does not inﬂuence denitriﬁcation rate. Data were analyzed by
ANOVA (F 4.10 = 140.1, P < 0.0001). Rates with different letters after symbols are

different based on Tukey’s HSD test (minimum signiﬁcant difference: 0.16).

18

activity, I also included control treatments in which phosphate buffer was added to soil
from the successional ﬁeld (pH 5.92 i 0.17; mean : SE) and calcium carbonate was
added to soil from the agricultural ﬁeld (pH 7.22 i 0.02; mean 1 SE). I measured pH in
three separate, replicate vials for all soils after inducing nos, which is the same point at

which I measured enzyme activity in the assay vials.

Nitriﬁer N 20 production

Nitrifying bacteria, which are obligately aerobic, are also an important source of
N20 in many soils (e. g. Bremner and Blackrner 197 8, Firestone and Davidson 1989).
Nitriﬁers tend to produce N20 at higher oxygen partial pressures than do denitriﬁers (e. g.
Davidson et al. 1986, Klemmedtsson et al. 1988b) but the oxygen threshold for nitriﬁer
N20 production is not well characterized (e. g. Goreau et al. 1980).

I investigated whether nitriﬁers contributed to N20 production at the very low
oxygen concentrations used in the enzyme activity assay by conducting incubations
similar to those described for enzyme induction. I incubated two sets of vials under
anaerobic conditions, one set with 10 percent acetylene and one set with no acetylene in
the headspace. I measured N20 production for about 8 h and then added 0.3 mL oxygen
(equilibrium solution concentration: 0.86 umol ' L'l) to both sets of incubation vials to
create microacrobic conditions. I continued to measure N20 production for an additional
8 h.

I was able to assess whether there was any nitriﬁer activity under the rrricroaerobic
conditions by ﬁrst calculating the ratio of N20 production rates under anaerobic versus
microacrobic conditions for both sets of vials. In a 10 percent acetylene atmosphere,
nitriﬁers are inhibited (Berg et al. 1982, Klemmedtsson et al. 1988a) regardless of the

oxygen level. Under an acetylene atmosphere, then, the ratio between N20 production

19

under anaerobic and microaerobic conditions is a measure of denitriﬁer inhibition by
oxygen. When no acetylene is present, nitriﬁers produce N20 only under microacrobic
conditions, if at all. If nitriﬁers contribute to N20 production, the anaerobic:microaerobic
N20 production ratio would be smaller than that in vials containing acetylene. If, on the
other hand, only denitriﬁers produce N20 under microacrobic conditions with no
acetylene, the anaerobic:microaerobic N20 production ratio should be of similar
magnitude in the presence or absence of acetylene. By comparing ratios in the presence
and absence of acetylene, then, I was able to determine whether there was nitriﬁer

activity in these incubations. I compared ratios using Student’s t test (Ott 1984).

Statistical Analyses

I used three separate three-way AN OVAs to determine the effects of site, pH, and
oxygen on net N20 production rate, denitriﬁcation rate and rNZO. These analyses were
conducted using the GLM procedure of SAS version 6.09 (SAS Institute 1996). I used
the Type III sum of squares due to the unbalanced design (Potvin 1993) and used the
LSMEAN S/PDIF F (least squares means) procedure to make a priori pairwise
comparisons when main effects were signiﬁcant. For rNZO I subjected data to an
arcsin(square root) transformation since values were constrained between 0 and 1

(Hinkelman and Kempthome 1994).

20

RESULTS
Calcium carbonate did not have any inﬂuence on erO for the soil from the
agricultural ﬁeld (ANOVA, F132 = 0.33 , P = 0.57) and phosphate buffer had no inﬂuence
on erO for the soil from the successional ﬁeld (ANOVA, F 1.14 = 0.19 , P = 0.67). For
clarity, I therefore present only results for soils incubated at native pH and at a pH level
adjusted to that of the other soil. I also present data from only two replicates of the soil
from the successional ﬁeld since one replicate soil sample from this ﬁeld was lost prior to

completing all enzyme assays.

Nitriﬁer N 20 production
The anaerobic:microaerobic N20 production rate ratios used to test for nitriﬁer
N20 production are provided in Table 2.2; there were no signiﬁcant differences in ratios

between vials with and without acetylene.

Preincubation: Denitriﬁer community enzyme induction assay

I assessed nitrous oxide reductase (nos) activity during induction assays by
comparing N20 accumulation curves for soils in the presence and absence of acetylene
(Figure 2.1). N20 production and consumption patterns similar to these have been
observed in a variety of soils (e.g. Klemmedtsson et al. 1988b, Dendooven and Anderson
1994) and previous studies, using both chlorarnphenicol and 13N, have established that
these patterns are due to the sequential induction of denitriﬁcation enzymes (Firestone
and Tiedje 1979, Smith et al. 1978, Smith and Tiedje 1979, Firestone and Tiedje 1979).
In general, the sequence of events is as follows: 1) N20 appears, indicating nar, nir and
nor activity, 2) N20 production rate increases, indicating de novo synthesis of additional

nar, nir, and nor, 3) similar N20 accumulation rates in the presence and absence of

21

Table 2.2. Nitrous oxide production rates with and without acetylene under anaerobic
and microacrobic conditions, and the ratio between anaerobic and microacrobic rates for
soils ﬁ'om the conventionally-tilled agricultural ﬁeld and the never-tilled successional

ﬁeld at the KBS LTER site.

 

Rates or rgtios Agricultural ﬁeld Successional ﬁeld

 

N20 production rate:
Without acetylene
Anaerobic 110 i 24 268 i 92
Microaerobic 46 i 12 127 i 47
With acetylene
Anaerobic 441 i 90 536 i 120
Microaerobic 170 i 36 248 i 63
Anaerobiczmicroaerobic N20
production rate ratio
Without acetylene 2.43 i 0.10a 2.13 : 0.06b
With acetylene - 2.64 i 0.192‘ 2.18 _+_ 0.07b

 

Note: Microaerobic conditions were created by adding 0.3 mL oxygen (equilibrium
solution concentration of 0.86 umol ' L'l) to previously anaerobic vials. Results show
that nitriﬁers do not produce N20 at low oxygen concentrations in these soils. Values are
means i 1 SE (n = 3 for agricultural ﬁeld; n = 2 for successional ﬁeld).

a (t test, taoj = 0.96, df= 3, P > 0.1).

b (t test, :00, = 0.54, df= 2, P > 0.1).

22 1

Figure 2.1. Denitriﬁer community enzyme induction curves at two pH levels for soils
from the conventionally-tilled agricultural ﬁeld and the never-tilled successional ﬁeld at
the KBS LTER site. Nitrous oxide reductase activity is indicated by greater N20
production in the presence than in the absence of acetylene. Values are means and error

bars are 3; 1 SE (n = 3 for agricultural ﬁeld; 11 = 2 for successional ﬁeld). Note different

ranges for x axes.

 

 

23

 

       

 

 

 

 

Eva_._.

on mm on m? or m o
e o
1m
row
in?
tow
892636.016 Fm

Emu 6.86385 .0
tom

Eves:

om on om om ov om ow o_. o
_ _ _ _ _ _ O
1m
to?
1a..
row

€22.63 3 In

26: .9238? .m ma
non

(t-nos B-o‘N 611)
uogronpOJd OZN angrelnwng

( 1-1108 B-o‘N 611)
uononpord OZN angrelnuino

Ev 65:

ommvovmmommwommwov m o

— L _ _ _ _ _ _ _

    

     
    
 

A6255 3 In
Em... 3:23.805 .0

EV 65:

ommvovmmommmommworm o
7. _ _ _ _ __

 

203.com $3 0
653.com o: O

6255 me In
26: 3.2.8:? .<

rm

10..
imw
ION
1mm
tom

 

 
        

In

low
19.
row
1mm
100

 

611)

N

D

(14108 oz
uogronpord OZN engrelnuino

(i-uos 5-OZN 611)
uogronpord OZN aAgrelnurnQ

Figure 2.1

24

acetylene indicate that nos either has not yet been induced or is not active, and 4) greater
N20 accumulation in the presence of acetylene than in its absence indicates that nos has
been induced and is active since N20 is being consumed.

In the soil from the agricultural ﬁeld N20 accumulation was greater in the
presence of acetylene than in its absence, indicating that there was nos activity less than
11 h after the beginning of the incubations (Figure 2.1A). When the pH of this soil was
adjusted to 5.7, N20 production in the presence of acetylene (nar, nir and nor induction)
and N20 consumption in the absence of acetylene (nos induction) were both greatly
inhibited (Figure 2.1B). Signiﬁcant denitriﬁcation did not occur before about 37 h and
N20 consumption in the absence of acetylene was inhibited for about 48 h. At both pH
levels, total denitriﬁcation was 23-25 pg N20 ' g soil'l.

In the soil from the successional ﬁeld, there was no nos activity at the beginning
of the incubations at either pH level (Figures 1C and D). Increasing soil pH to 6.9
resulted in earlier production and consumption of N20, indicating that low pH had an
inhibitory effect on nos activity in this soil too. Total denitriﬁcation was about 27 pg
N20 ' g soil'1 at both pH levels.

In a separate set of preincubation vials, I observed a positive, linear relationship
between the amount of nitrate added to slurries and total denitriﬁcation (Table 2.3),
suggesting that total denitriﬁcation was limited by nitrate depletion in these soils. In
addition, vigorous denitriﬁcation resumed when additional nitrate was added to slurries
after N20 accrunulation had ceased (data not shown). Since N20 production in the
presence of acetylene is stoichiometrically related to nitrate consumption during
denitriﬁcation, I was able to calculate the approximate amount of nitrate present in soil

slurries at each point along the induction curves based on headspace N20 concentrations.

25

Table 2.3. Total denitriﬁcation for soil from the agricultural ﬁeld at the KBS LTER site

with increasing rate of nitrate addition.

 

 

 

 

NO3-N added Total N20 production
(pygmy-1 (112. N;Q=_gsoﬂi)
0 12.7 i 0.04
16.7 38.4 i 2.05
33.3 50.1
50.0 62.1 i 2.19
66.7 87.2 i 0.57

 

Note: Soil was incubated anaerobically with 10 percent acetylene in the headspace.
Values are means i 1 SE (n = 2, except for when 33 pg NO3-N was added ' g soil'l, in
which case n = 1). The regression equation of N20 production vs nitrate is y = 1.04x +

15.6, r = 0.99. Results show that denitriﬁcation was nitrate limited in unamended slurries.

26

These calculations showed that nitrate concentration at the approximate time of initial nos
activity was lower in the soil from the agricultural ﬁeld than in the soil from the
successional ﬁeld at all pH values (Table 2.4), and that nitrate concentration at the time of

nos induction increased with increased pH.

Denitriﬁer community enzyme activity assay

Shaking soil slunies at 400 rpm did not kill cells: aerobic heterotrophic bacterial
counts were 2.9 i 0.7 x 10‘5 colony forming units (CFU) ° g soil'l after gentle shaking and
3.5 i 0.5 x 106 CFU ' g soil'l (means i SE) following vigorous shaking (t test, t0.05.1 =
409, IF 3,P>O.l). ‘

Site, pH, and oxygen each affected denitriﬁcation rate (Figure 2.2A, B; Table
2.5). Denitriﬁcation rate was higher in the soil from the successional ﬁeld than in the soil
from the agricultural ﬁeld at either pH level (P 5 0.0001, pairwise comparisons between
two soils at each oxygen level). A pH effect was observed only for the soil from the
successional ﬁeld (Figure 2.2A, B; P 5 0.0001 for soil from the successional ﬁeld; P >
0.1 for soil from the agricultural ﬁeld). Denitriﬁcation rates decreased exponentially with
increasing oxygen concentration for both sites at both pH levels (Figure 2.2A, B), but the
rate of decrease was different for the two sites. For the soil from the agricultural ﬁeld,
there was a signiﬁcant decrease in denitriﬁcation rate at both pH levels only between 0
and 0.05 mL oxygen (P < 0.005). For the soil from the successional ﬁeld, there was a
signiﬁcant decrease in denitriﬁcation rate at both pH levels between 0 and 0.05 mL
oxygen (P < 0.0005) and also between 0.05 and 0.10 mL oxygen at pH 5.7 (P < 0.01).

Site, pH, and oxygen each affected net N20 production rate (Figure 2.2C, D;

Table 2.5), but the relationships were sometimes complicated. At pH 5.7, net N20

27

Table 2.4. Calculated slurry nitrate concentrations at two pH levels at the point of initial

nos activity during preincubations for soils from the conventionally-tilled agricultural

ﬁeld and the never-tilled successional ﬁeld at the KBS LTER site.

 

Nitrate concentration

 

Field pH (62 N01ﬁ= soili)
Agricultural 5.7 1.61 i 0.53
6.6 (native) 3.47 i 0.42
Successional 5.7 (native) 6.90 i 0.77
6.9 10.80i1.10

 

Note: Values are means i 1 SE (n = 3 for agricultural ﬁeld; n = 2 for successional ﬁeld).

28

Figure 2.2. Denitriﬁcation (A and B) and N20 production rates (C and D) at four oxygen
concentrations and two pH levels for soils from the conventionally-tilled agricultural ﬁeld
and the never-tilled successional ﬁeld at the KBS LTER site. Values are means and error

bars are i 1 SE (n = 3 for the agricultural ﬁeld; n = 2 for the successional ﬁeld).

29

eouvm com>xo 1::
2.6 one mod 0

e\o\b1\m

me In 0
82655.0. :6 o

20: _mco_ww0003w .D

emcee comes 1.6
de owd 3.0 my

10

row
tom
row
row

 

10m

 

me :a c

82.65 3 :6 0
Bo: 6:23.805 .m

nor
10m
low
low

 

rom

(191w 1-uos 5-OZN Bu) '

(1-U!w1-1!os 5-OZN 6U)
are; uogronpord ozN raw

9121 uogreoguriuea

venue com>xo 1::
mmd 9.6 mo_.o w

 

 

 

o
1 N
1 v
8265 me In 0 1 m
3 In 0 1 m
26: 5336:? .o 1 e
umuum comes 1::
8d 03 mod o
_ _ _. _ O
N
1 v
1 o
3255 me In 0 m
3 In 0 1 2
22 3.3.8:? .<
r we

 

(1-urw1-uos 6-ozN 5U)
9191 uouonpord QZN reN

(1-U!w 1110s 5-OZN 6U)

9131 uoneouuriuac]

Figure 2.2

30

Table 2.5. F values of three-way ANOVAs to determine effects of site, pH, and oxygen

on denitriﬁcation rate; net N20 production rate; and rNZO, the relative rate of N20

production for soils from the conventionally-tilled agricultural ﬁeld and the never-tilled

successional ﬁeld at the KBS LTER site.

 

 

 

Source of
Variable vaL'ation df MS F P
Denitriﬁcation Site 1 3450.1 924.4 0.0001
rate pH 1 323.8 86.8 0.0001
Oxygen 3 200.4 53.7 0.0001
Soil x pH 1 378.5 101.4 0.0001
Oxygen x soil 3 20.7 5.6 0.0049
Oxygen x pH 3 1.7 0.5 0.71
Oxygen x soil x pH 3 1.4 0.4 0.78
Net N20 Site 1 104.1 163.8 0.0001
rate pH 1 87.9 138.3 0.0001
production Oxygen 3 3.4 5.4 0.0057
rate Soil x pH 1 38.2 60.2 0.0001
Oxygen x soil 3 13.5 21.2 0. 0001
Oxygen x pH 3 0.1 0.2 0.91
Oxygen x soil x pH 3 3.5 5.5 0.0054
rNZOa Site 1 1.162 244 5 0.0001
pH 1 0.550 115 7 0.0001
Oxygen 3 0.235 49 4 0.0001
Soil x pH 1 0.117 24 6 0.0001
Oxygen x soil 3 0.007 1 4 0.26
Oxygen x pH 3 0.002 0.4 0.73
Oxygen x soil x pH 3 0. 001 0.3 0. 86

 

Note: n= 3 for agricultural ﬁeld, n= 2 for successional ﬁeld.
data were subjected to an arcsin(square root) transformation since values were
constrained between 0 and 1.

31

production rate was higher in the soil from the successional ﬁeld than in the soil from the
agricultural ﬁeld (P 5 0.0001) at all oxygen levels except under anaerobic conditions (P >
0.1). Net N20 production rate in the soil from the successional ﬁeld at pH 6.9 was not
different than for the soil from the agricultural ﬁeld at either pH level (P > 0.1). For the
soil from the agricultural ﬁeld net N20 production rate was similar at both pH levels (P >
0.1) except under anaerobic conditions, in which case the rate was much higher at pH 5.7
than 6.6 (P 5 0.0001). Net N20 production rate also did not change with oxygen level at
this site (P > 0.1) except for the decrease between 0 and 0.05 mL oxygen at pH 5.7 (P <
0.0005). For the soil from the successional ﬁeld net N20 production rate was higher at
native pH than at pH 6.9 at each oxygen level (P 5 0.0001). At both pH levels, net NZO
production rate increased with oxygen (P 5 0.0001, pairwise comparison between 0 and
0.15 mL oxygen).

The rNZO increased with increasing oxygen for both sites at both pH levels
(Figure 2.3; Table 2.5; P _<_ 0.0001, pairwise comparison between 0 and 0.15 mL oxygen).
The increase tended to be exponential for soil from the successional ﬁeld and logistic for
the soil from the agricultural ﬁeld regardless of pH level. There were no signiﬁcant
interactions between oxygen and site, oxygen and pH, nor oxygen, site and pH (Table
2.5). A signiﬁcant interaction between site and pH (Table 2.5), however, is illustrated in
Figure 2.3: there was a larger change in rNZO when slurry pH was adjusted for the soil
from the agricultural ﬁeld than for the soil from the successional ﬁeld (P _<_ 0.0001). In
addition, the two sites differed in rNZO whether compared at pH 5.7 (P 5 0.0001) or 6.6-
6.9 (P < 0.011), and this difference was independent of oxygen level. At native pH, there
was no signiﬁcant difference in rNZO under anaerobic conditions (P > 0.1), a slightly
signiﬁcant difference at 0.05 mL oxygen (P < 0.05) and barely non-signiﬁcant
differences at 0.10 (P < 0.056) and 0.15 mL oxygen(P < 0.064).

32

Figure 2.3. Relative rate of N20 accumulation, rNZO, at four oxygen and two pH levels
for soil from the conventionally-tilled agricultural ﬁeld and the never-tilled successional
ﬁeld at the KBS LTER site. Values are means and error bars are i 1 SE (n = 3 for the

agricultural ﬁeld; 11 = 2 for the successional ﬁeld).

33

 

 

 

1 q
8 0.9 - . Agricultural ﬁeld
§ 0.8- pH 5.7
E. 3" 0-7 ' I Agricultural ﬁeld
of?” 0.5.. pH 6.6 (native)
2 z
‘5 § 0-5' O Successional ﬁeld
E ‘2?» 0.4- pH 5.7 (native)
6: <1
2% _ 0'3‘ Cl Successional ﬁeld
a 0.2 - pH 6.9
1:

0.1 -
0 T

I 1 I
O 0.05 0.10 0.15

mL oxygen added

Figure 2.3

DISCUSSION
Enzyme activity assays

By controlling all known environmental regulators of denitriﬁcation rate and
erO (temperature, moisture, enzyme induction status, oxygen, pH), or by providing
them in non-limiting amounts and without diffusion limitation (carbon, nitrate), I created
incubation conditions in which the only variable inﬂuencing denitriﬁcation rate (Figures
2A, B) and erO (Figure 2.3) at a given oxygen concentration and pH level was the
composition of the denitrifying community.

The denitrifying community in the soil from the agricultural ﬁeld had enzymes
involved in N20 production (nar, nir, nor) that were more sensitive to oxygen than that I
ﬁom the successional ﬁeld, as indicated by the greater rate of decrease in denitriﬁcation
rate with increasing oxygen in the soil from the agricultural ﬁeld (Figure 2.2A, B). The
denitrifying community in the soil from the successional ﬁeld, on the other hand, had
enzymes involved in N20 production that were more sensitive to pH than were those
from the community from the agricultural ﬁeld, since denitriﬁcation rate was greater in
this community at native than elevated pH (Figure 2.2B). This pH-sensitivity is
consistent with the concept that the pH optima of soil denitriﬁer communities (Parkin et
al. 1985) and isolates (Burth and Ottow 1983) reﬂect the pH of their native environment.
In contrast, the denitriﬁer community in the soil from the agricultural ﬁeld seems to have
a broader pH optimum since total denitriﬁcation was the same at both low and high pH
levels (Figure 2.2A).

The sensitivity of nos activity to oxygen was different in these two soil
denitrifying communities as indicated by differences in the shape of their erO curves,
regardless of pH (Figure 2.3). In addition, rNZO was different for the two soils at their

native pH (Figure 2.3). This difference in rNZO was not due to pH; if it had been a pH

35

effect, the soil from the successional ﬁeld would have had a higher rNZO than the soil
from the agricultural ﬁeld since nos activity is generally inhibited by low pH (e. g.
Nomrrrik 1956, Focht 1974, Blackmer and Bremner 1978, Firestone et al. 1979, 1980,
Koskinen and Keeney 1982, Weier and Gilliam 1986, Christensen and Tiedje 1988,
Christensen et al. 1990b). That the soil from the successional ﬁeld always had a
signiﬁcantly lower erO than the soil from the agricultural ﬁeld when incubated at
similar pH (Figure 2.3) conﬁrms that differences in rNZO were not due to pH. Since all
environmental regulators of erO are controlled or provided in non-limiting amounts in
these incubations, differences in rNZO must be due to differences between the two
denitriﬁer communities in nos enzyme sensitivity to oxygen.

Since the two denitriﬁer communities respond differently to the same
environmental variables I conclude that they must be comprised of different denitriﬁer
taxa. Indeed, the denitriﬁer community structure, based on denitrifying bacteria isolated
from these same soil samples, was very different between soils (Chapter 2). It follows,
then, that in situ N20 production is inﬂuenced by denitriﬁer community composition in
these two soils -- under identical environmental conditions a greater proportion of total
denitriﬁcation ﬂux (N 20 + N2) remained as N20 in the soil from the agricultural ﬁeld
than in the soil ﬁom the successional ﬁeld. These results suggest a signiﬁcant functional

role for soil denitriﬁer community composition in these two soils.

Nitriﬁer N 20 production
The lack of signiﬁcant difference between anaerobic:microaerobic N20
production rate ratios (Table 2.2) indicates that nitriﬁers were not a signiﬁcant source of
N20 in these incubations. Denitriﬁcation, then, is the source of all N20 production in

these assays.

36

Preincubation: enzyme induction assays

Although not the primary focus of this study, I found evidence that the regulation
of denitriﬁcation enzyme induction may also differ between these two denitrifying
commrmities. Nitrate and/or nitrite concentrations have been shown to be important
regulators of nos induction (Blackmer and Bremner 1978, 1979, Firestone et al. 1979,
1980). In the preincubations, nitrate concentration at the time of initial nos activity was
different for the two denitrifying communities (Table 2.4). This difference was not a pH
effect since the difference in nitrate concentration at the time of initial nos activity was
greater between the two soils when preincubated at similar pH than when preincubated at
native pH (Table 2.4). These differences may also be due to differences in denitriﬁer
community composition, but since I did not control all environmental variables important
in regulating denitriﬁcation enzyme induction in the preincubations, further work is
required to conﬁrm this conclusion.

The denitrifying communities in the two soils were similar in some N20
production characteristics. Nitrous oxide reductase activity appeared after 13 to 55 h in
all soils regardless of pH level (Figure 2.1), a period consistent with the 16-33 h threshold
observed by others (Firestone and Tiedje 1979, Dendooven and Anderson 1994,
Dendooven et al. 1996) for a number of different soils. Also, initial denitriﬁer enzyme
activity seemed to be inhibited at pH 5.7 in both communities (Figure 2.1).

The presence of nos activity in soil from the agricultural ﬁeld at the beginning of
induction incubations (Figure 2.1) suggests that there were anaerobic microsites in this
soil despite very dry conditions at the time of sampling (Table 1.1). Or, it may be that
denitriﬁcation enzymes were maintained in the ﬁeld under non-denitrifying conditions

(Smith and Parsons 1985).

37

Significance

My ﬁndings provide evidence that microbial diversity is not necessarily
ﬁmctionally redundant, as often assumed. Though redundancy may be the case within
such broad ﬁrnctional groups as carbon mineralizers (e. g. Schimel 1995), my results
support the developing view that diversity is likely to be ﬁrnctionally signiﬁcant within
more narrowly-deﬁned functional groups such as the denitriﬁers (Conrad 1996, Schimel
1995). These results also have important practical implications. Efforts to model N20
ﬂux from soils to the atmosphere may need to incorporate some measure of denitriﬁer
community composition in order to accurately predict N20 ﬂuxes. Could it be that some
agricultural practices such as fertilization, tillage and/or plant community manipulation I
select for denitrifying organisms with lower nos activity than in adjacent, less-intensively
managed ﬁelds? If so, land use policy changes that are likely to emerge from recent
climate conventions (Bolin 1998) may need to consider, for example, the effect of

agricultural practices on the selection of denitriﬁer communities.

38

CONCLUSIONS

My results support the hypothesis that denitriﬁer community composition
inﬂuences N20 production in soils. I collected samples from two geomorphically similar
soils that differed in plant community composition and disturbance regime -- an
agricultural ﬁeld and a successional ﬁeld. By controlling, or providing in non-limiting
amounts, all known environmental regulators of denitriﬁer N20 production and
consumption, I created conditions in which the only variable contributing to differences
in denitriﬁcation rate and erO in the two soils was the denitriﬁer community. Both
denitriﬁcation rate and rNZO differed for the two soils under these controlled incubation
conditions. Oxygen inhibited the activity of enzymes involved in N20 production (nar, '
nir, nor) to a greater extent in the denitrifying community from the agricultural ﬁeld than
that from the successional ﬁeld. The nar, nir and nor enzymes of the denitrifying
community from the successional ﬁeld, on the other hand, were more sensitive to pH than
were those in the denitrifying community from the agricultural ﬁeld. Under identical
environmental conditions, the denitrifying community in the soil from the successional
ﬁeld had relatively more active nos enzymes, which reduce N20 to N2, than the
denitrifying community in the agricultural ﬁeld. Also, the shape of the erO curve with
increasing oxygen was different for each denitrifying community. Each of these
differences suggests that the denitrifying communities in these two soils are different and
that they do not respond to environmental regulators in the same manner.

I am not aware of other studies showing that native microbial community
composition can regulate an important ecosystem function. Models of N20 ﬂux from
soils may need to include the inﬂuence of microbial community composition on N20

ﬂux.

Chapter 3
THE ROLE OF BIOTIC DIVERSITY IN RATES OF NITROUS OXIDE
CONSUMPTION IN A TERRESTRIAL ECOSYSTEM

INTRODUCTION

Microbial diversity is usually assumed to be functionally redundant such that
ecosystem functions mediated by microbes are usually considered to be controlled solely
by abiotic factors and trophic-level interactions (e.g. Meyer 1993, Beare et al. 1995, Heal
et al. 1996). Thus, in both mathematical and schematic models of biogeocherrrical
processes, microbial communities are usually treated as black boxes that transform inputs
to outputs at rates deﬁned by environmental parameters empirically calibrated. For
general processes such as carbon turnover these models appear to work well, thereby
implicitly supporting the assumption that microbial diversity is functionally redundant
(e.g. Schimel 1995).

For other biogeochemical processes, however, predicting in situ ﬂux rates has
proven difﬁcult. In situ N20 ﬂux rates, for example, have proven very difﬁcult to predict
despite the incorporation of substantial environmental detail into N20 models (e. g.
McGill et al. 1981; McConnaughey and Bouldin 1985; Rolston et al. 1984; Johnsson et
a1. 1987; Parton et al. 1988; Arah and Smith 1989; Li et al. 1992a, b; Ojima et al. 1992;
Smith et al. 1993; Parton et al. 1995). This lack of predictive ability may be because
differences in microbial communities are not incorporated into contemporary models. If
communities vary in their taxonomic makeup, and if different denitrifying taxa produce

or consume N20 at different rates under the same environmental conditions, then

39

40

prediction of N20 ﬂux will be difﬁcult without incorporating these differences into
models.

Modelling N20 ﬂux is important because the atmospheric concentration of this
important greenhouse gas and natural catalyst of stratospheric ozone degradation is
increasing (e. g. Houghton et al. 1996) and models are needed to predict further effects of
disturbance, and to evaluate mitigation efforts.

Soil is a major source of atmospheric N20 and denitrifying and nitrifying bacteria
are the primary biological sources of N20. Denitriﬁers are the only organisms that
consume N20 (Conrad 1996). Since denitriﬁers are one of the best-studied and most
phylogenetically diverse groups of soil microorganisms (Tiedje 1994, Zumft 1992, 1997),
they also provide a good model for studying the inﬂuence of microbial diversity on
ecosystem ﬁrnction.

Nitrous oxide production by denitriﬁers reﬂects the relative activity of nitric oxide
reductase (nor) and nitrous oxide reductase (nos; Betlach and Tiedje 1981), the enzymes
responsible for the last two steps in the denitriﬁcation pathway (Tiedje 1994):

NO3' —) N02' —> NO —-> N20 —-) N2.
nar nir nor nos
The ﬁrst two enzymes in the process are nitrate reductase (nar) and nitrite reductase (nir).
The four denitriﬁcation enzymes are usually induced sequentially under anaerobic
conditions and oxygen partial pressure, carbonznitrate substrate ratio, and pH are the
primary environmental regulators of denitriﬁer enzyme synthesis and activity (e. g. Tiedje
1988).

Regulation of denitriﬁcation enzymes by these environmental variables seems to

vary considerably among denitrifying isolates. For example, in a series of studies

comparing regulation of N20 production among select denitriﬁers, N20 production and

41

the NzOzNz ratio were inﬂuenced more by bacterial species than by oxygen partial
pressure (Abou Seada and Ottow 1985, Anderson and Levine 1986, Munch 1991), soil
texture (and its effect on oxygen difﬁrsion; Munch and Ottow 1986), nitrate concentration
(Munch 1989), or pH (Burth and Ottow 1983). In addition, three denitriﬁers grown under
the same conditions differed not only in grth rate and cell yield, but also in the ability
to utilize exogenous denitriﬁcation intermediates and in the relative rates of various steps
in the denitriﬁcation pathway (Carlson and Ingraham 1983). More recently, nitrite
reductase regulation by oxygen, nitrate, and nitrite was shown to differ, sometimes
dramatically, among seven denitriﬁer strains commonly found in soils (Ka et al. 1997).
This physiological diversity among select denitrifying isolates suggests that denitriﬁer
community structure may inﬂuence ecosystem ﬁmction.

No researchers to date, however, have measured denitriﬁcation enzyme activity
under controlled conditions among organisms isolated ﬁ'om the same natural community.
The studies cited above used small numbers of denitrifying isolates, usually chosen from
culture collections. The few studies that include some measure of both native denitriﬁer
community composition and denitriﬁcation enzyme activity have only measured the end
products of denitriﬁcation at one point in time (Gamble et a1. 1977, Christenson and
Bonde 1985, Weier and MacRae 1992), i.e. they have not compared the isolates’
denitriﬁcation enzyme sensitivities to environmental regulators. If physiological
differences among isolates are to be related to community-level N20 production
potentials, then such differences must be demonstrated among organisms belonging to the
same native communities.

In Chapter 2 I present evidence, based on soil denitriﬁcation enzyme assays, that
soil denitriﬁer community structure inﬂuences potential N20 production in soils. In this

chapter I ﬁrrther this investigation of the role of biotic diversity on potential N20

42

production in soils by focusing on the physiology of the denitrifying bacteria involved in
N20 production and consumption. Speciﬁcally, I quantify the sensitivity of nitrous oxide
reductase (nos) activity to oxygen among denitriﬁers isolated from two geomorphically
similar soils from ﬁelds that differed in plant community composition and disturbance
regime. I hypothesize that community structure will be different between these two sites
and that there will be physiological diversity in nos sensitivity to oxygen among isolates,
thus demonstrating a functional signiﬁcance to taxonomic diversity among denitriﬁers.

I chose to study the regulation of nos activity rather than other denitriﬁcation
enzymes because nos is unique to denitriﬁers, its activity is important to N20 ﬂuxes in
situ (e. g. Firestone and Davidson 1989, Dendooven and Anderson 1994) and thus to
global N20 ﬂuxes, and nos activity has received little attention compared to that of other

denitriﬁcation enzymes (e. g. Hochstein and Tomlinson 1988).

43

MATERIALS AND METHODS
Study Site and Soils
The study site and soils are described in Chapter 1, including Table 1.1. I used
the same soil samples in the research described in both this chapter and Chapter 2. After
seiving, soil samples were stored at 4°C until subsampled for plate counts and isolations,

both of which were initiated within 10 days.

Media and growth conditions

Throughout this study I used a modiﬁed R2A medium (R2A*) for both liquid and
solid media and grew all isolates in an anaerobic glove box (Coy, Ann Arbor, MI) with an
atmosphere of 5 percent C02, 10 percent H2 and the balance N2. The R2A* contained,
per liter: 1.7 g NH4C1, 4.0 g KNO3, 2.8 g sodium succinate, 1.5 g soditun pyruvate, 2.0 g
peptone, 2.0 g casarnino acids, 2.0 g yeast extract, 1.0 g KHZPO4, 1.6 g KZHPO4, 0.06 g
CaC12 ' 2HZO, 0.08 g MgClz ' 6HZO, 0.028 g F eSO4 ' 7HZO, 0.02 g NaZSO4, 0.04 g MnClz
' 4HZO, 0.004 g H3BO3, 0.004 g ZnClz, 0.004 g NiC12 ' 6HZO, 0.0012 g CuClz ' 2HZO,
and 0.0004 g NazMoO4 (Fries et al. 1994). No vitamins were included. The ﬁnal solution
was adjusted to pH 7.2 using NaOH. Solid media was made by adding 15 g Bacto
(Difco, Detroit, MI) agar per liter. I monitored anaerobic conditions in the glove box
using a bottle of sterile meditun containing resazurin dye and adjusted to -210 Eh using
cysteine ' HCl (Costilow 1981), and by periodically sampling the glove box atmosphere
for oxygen using a gas-tight needle and syringe. Gas samples were analyzed using an HP
5890H GC (Hewlett Packard, Rolling Meadows, IL) equipped with a Poropak Q column
and a 63Ni electron capture detector (ECD). Bacterial transfers were made under aerobic
conditions inside a laminar ﬂow hood; denitrifying bacteria seem to grow best when

initially exposed to air (R. E. Murray, personal communication).

44

Enumeration and isolation

I enumerated and isolated denitrifying bacteria using techniques slightly modiﬁed
from those of Gamble et al. (1977). I prepared a 10-fold dilution series by adding 10 g
soil (wet weight) and 2-3 drops Tween 80 to 90 mL phosphate buffer (pH 7.0). I used a
Waring blender to mix the soil (speed 5) for two minutes and then pippetted, while
stirring, 1 mL of solution into 9 mL of phosphate buffer (pH 7.0). R2A* agar plates were
inoculated with 0.1 mL of solution from the 10'2 to 10'5 dilution tubes. After 3 to 6 d
growth in the anaerobic glove box I counted all colonies on plates containing feWer than
300 colonies (10‘1 - 10'6 dilution plates). I then transferred all colonies from one replicate
of each of these plates to test tubes containing R2A* broth and a small inverted tube to '
capture gases formed during growth. I disposed of those colonies not producing gas.
Gas-producers (potential denitriﬁers) were streak-plated until pure. Puriﬁed isolates were
conﬁrmed as denitriﬁers if they converted at least 80 percent of the nitrate (5 mM) in 5
mL of R2A* to N20 when grown anaerobically under a 10 percent (v/v) acetylene
headspace (Mahne and Tiedje 1995). Acetylene blocks the reduction of N20 to N2 by nos
(Balderston et al. 1976, Yoshinari and Knowles 1976); the resulting N20 is more easily
measured than N2. The number of viable anaerobic heterotrophic bacteria, gas-producers,
and denitriﬁers in each soil was calculated using standard counting techniques (e. g.
Gamble et al. 1977, Zuberer 1994). Puriﬁed isolates were grown anaerobically for 36-48
h in R2A* broth and stored in a 50:50 solution of sterile glycerol and R2A* broth at -
20°C.

Isolate characterization
I characterized denitrifying isolates using fatty acid methyl ester (FAME) proﬁle

analysis. I grew isolates on R2A* agar for 96 i 4 h at 27-28°C in the anaerobic glove

45

box. I harvested the entire plate using loops formed from glass pipettes. The harvested
biomass was stored in ashed, capped test tubes at -20°C until fatty acids were extracted. I
used the Microbial Identiﬁcation System (MIDI, Inc., Newark, DE) protocol for fatty acid
extraction and analysis (Sasser 1990) with only slight modiﬁcations. Brieﬂy, lipids were
saponiﬁed using hot NaOH in methanol, and methylated by adding hot HCl in methanol.
F AMEs were extracted from this solution using methyl tert-butyl ether (MTBE) in
hexane and washed using dilute NaOH. I then transferred the organic phase containing
the FAMEs to a new ashed test tube, concentrated the FAMEs by evaporating the MTBE
solvent under a stream of N2 gas, and then reconstituted FAMEs by adding 100 pL of
MTBE solvent. I transferrred this solution to a GC vial for subsequent analysis by gas- 1
liquid chromatography using a Hewlett Packard 5890 GC equipped with an Ultra 2
capillary column (crosslinked 5% Ph Me silicone, 25 m x 0.2 mm x 0.33 mm ﬁhn
thickness) and a ﬂame ionization detector. Each isolate was analyzed at least in
duplicate; a random subset of isolates was analyzed up to ﬁve times.

I used previously characterized denitrifying isolates collected from the lab of J .M.
Tiedje (Michigan State University; Table 3.1) as reference strains. Reference strains
represented nine of the 14 denitrifying species commonly recovered from soils, and two
of four genera recovered with moderate to low frequency (Tiedje 1994). Six ﬂuorescent
pseudomonads that do not denitrify -- four strains of Pseudomonas putida and two
biovars of P. fluorescens (A and G) -- were also included as reference strains. FAME
proﬁles of reference strains grown under identical conditions as test organisms (Table
A.1) were used to help identify the unknown isolates and to determine the level of
dissimilarity that deﬁned taxa using cluster analysis. In addition, two of these isolates
(Alcaligenes xylosoxidans subsp. xylosoxidans NCIB 11015 and P. stutzeri JM300) were

run as positive controls with each set of unknowns.

46

Table 3.1. Reference strains used to compare cellular fatty acid proﬁles against those of

denitrifying bacteria isolated from the conventionally-tilled agricultural ﬁeld and the

never-tilled successional ﬁeld at the KBS LTER site.

 

 

Phylogenetic group
(and proteobactera
subdivision) Species 8mg
Proteobacteria A grobacterium tumefaciens ° G41
(alpha subdivision) Rhizobium sp.c OK-55
Proteobacteria Alcaligenes xylosoxidans subsp. denitrificans ° G65
(beta subdivision) Alcaligenes xylosoxidans subsp. denitriﬁcans G191
Alcaligenes xylosoxidans subsp. xylosoxidans NCIB 1 1015
Achromobacter cycloclastes ATCC 21921
Pseudomonas type 11 G107
Pseudomonas type 11 G143
Pseudomonas type 11 G163
Pseudomonas type 11 G188
Proteobacteria Pseudomonas aeruginosa ATCC PAOl
(gamma subdivision, Pseudomonas aureofaciens ATCC 17415
rRNA group I) Pseudomonas aureofaciens ATCC 17417
Pseudomonas chlororaphis ATCC 17810
Pseudomonas chlororaphis ATCC 17812
Pseudomonasﬂuorescens ATCC 948
Pseudomonasﬂuorescens ATCC 33512
* Pseudomonasﬂuorescens bv. A ATCC 17552
Pseudomonasﬂuorescens bv. B ATCC 17467
Pseudomonasﬂuorescens bv. B ATCC 17812
Pseudomonasﬂuorescens bv. C ATCC 17406
Pseudomonasﬂuorescens bv. C ATCC 17561
Pseudomonasﬂuorescens bv. F ATCC 12983
Pseudomonasﬂuorescens bv. F ATCC 17513
* Pseudomonasﬂuorescens bv. G ATCC 17386
* Pseudomonas putida 39
* Pseudomonas putida ATCC 17472
* Pseudomonas putida ATCC 25571
* Pseudomonas putida A ATCC 17391
Pseudomonas stutzeri JM300
Proteobacteria Burkholderia pickettii ° PKOl
(gamma subdivision,
rRNA group 11)
Gram positive Bacillus sp. G192
Bacillus SD. G193

 

 

 

47

Table 3.1. (continued)

 

 

Phylogenetic group
(and proteobactera
subdivision) Species Streirf
Uncertain Aquaspirillum itersonii
classiﬁcation ° Pseudomonas unknown type 3 G108

Alcaligenesfaecalis G148

 

: Strains with a G designation are originally from Gamble et al. (1977).
Originally identiﬁed by Gamble et al. (1977) as Alcaligenesfaecalis; reclassiﬁed by
Ka et a1. (1997), using 16S rRNA sequencing.

° Originally identiﬁed by Gamble et al. (1977) as Pseudomonas sp.; reclassiﬁed by

d Ka et al. (1997), using 16S rRNA sequencing.
Previously identiﬁed as Pseudomonas picketii; all rRNA group II Pseudomonas have
been reclassiﬁed into the new genus Burkholderia (Y abuuchi et al. 1992, and Urakarni
et al. 1994 cited in Zumft 1997).

c I considered these isolates of uncertain classiﬁcation: Aquaspirillium itersonii is
classiﬁed in two different taxa by Stackebrandt (1992); G108 was classiﬁed as
Pseudomonas of unknown type by Gamble et al. (1977); and other denitriﬁers

identiﬁed by Gamble et a]. (1977) as Alcaligenesfaecalis have since been reclassiﬁed
by Ka et al. (1997).

* Not a denitriﬁer.

48

Statistical analyses of FAME data

Since the number of FAME peaks identiﬁed increases with the amount of biomass
extracted, I standardized GC peak sizes to that of the isolate with the lowest biomass (as
estimated by total FAMEs extracted). Individual FAME peaks that fell below the GC
detection limit (set at 500 peak height units) after standardization were dropped from the
data set. This precaution ensures that differences found among isolates are due to
inherent differences in cellular fatty acid content, and not to differences in the amount of
biomass harvested and extracted.

FAME proﬁles of known and unknown isolates were then subjected to
hierarchical cluster analysis using NTSYS (Applied Biostatistics, Inc., Setauket, NY). I
ran a preliminary cluster analysis using all isolate replicates as separate samples to assess
replicate similarity. In the few cases where replicates did not cluster together, I regrew
isolates and extracted and analyzed their FAME proﬁles to identify outliers in initial runs.
Average FAME proﬁles were then calculated for each isolate and these data were
subjected to cluster analysis. I constructed dendrograms using the Euclidean distance
metric and UPGMA (unweighted pair-groups method using arithmetic averages) linkage.
Previous cluster analyses of FAME data showed that a number of combinations of
distance metrics and linkage methods with biases toward the formation of different types
of clusters (Everitt 1980, Milligan and Cooper 1987) gave essentially identical

dendrograms (Cavigelli et al. 1995).

Nitrous oxide reductase activity assay
I quantiﬁed nos sensitivity to oxygen for select isolates by measuring N20

consumption rates under anaerobic conditions and at three separate, very low oxygen

49

concentrations. I selected one or two isolates from each taxa identiﬁed by cluster analysis
and all isolates from one taxa containing seven unknown isolates (taxon 30).

I ﬁrst conﬁrmed nos activity in the selected isolates by testing for N20
consumption. For each isolate I transferred three drops of cell culture at late log phase
(approx. 40 percent transmittance at 560 nm when grown aerobically) to six 38 mL vials
containing 3 mL of degassed R2A* and an anaerobic headspace. I added acetylene (3.6
mL) aseptically to three of these vials and shook all vials at 100 rpm. I monitored
headspace N20 using the same GC used to measure oxygen in glove box atmosphere
samples. After N20 production had appeared in vials containing acetylene (evidence that
the enzymes leading to the production of N20 are active), I injected N20 (5 mL of 1.01 '
percent standard, which resulted in 70 pmol N20 ' L'1 in solution) into the vials
containing no acetylene and monitored headspace N20; N20 consumption conﬁrmed nos
activity.

I used a second set of incubations to measure the effect of oxygen on isolate N20
consumption rates. I again transferred three drops of cell culture at late log phase, this
time to a serum vial (120 or 160 mL) containing anaerobic R2A* broth (40 or 50 mL,
respectively) and 5 mL of 1.01 percent N20. I monitored headspace N20 concentration
and, after N20 consumption began; I added about 10 psi N2 (UHP; 99.999 percent) to
create a positive internal pressure to help avoid introduction of oxygen when transferring
cultures. I transferred 2 mL aliquots of the cell culture to four autoclaved 38 mL serum
vials containing an anaerobic atmosphere and I equilibrated headspace pressure with
atmospheric pressure. I then added 0, 0.25, 0.50, or 0.75 mL of laboratory atmosphere
(ﬁltered through a 0.22 pm ﬁlter) and 5 mL of a 1.01 percent N20 standard to each vial.
Vials were shaken on an orbit shaker (Model 3520, Lab-Line Instruments, Melrose Park,

IL) at 400 rpm, a speed previously determined to be sufﬁcient for eliminating oxygen

50

diffusion effects on denitriﬁcation rate in 38 mL vials containing three g soil plus 3 mL
water (Chapter 2). I then measured N20 and oxygen every 7-10 minutes for a period not
longer than 120 min to determine isolate N20 consumption rates. I replicated this
procedure three to ﬁve times for each test isolate. I used published values of the Bunsen
coefﬁcient to calculate total N20 in vials and oxygen concentration in solution.

I tested for differences in N20 consumption rates at each oxygen level for each
isolate using analysis of covariance (oxygen as main effect and time as covariate) with
the GLM procedure in SAS, version 6.09 (SAS Institute 1996) and I compared
consumption rates at each oxygen level using the SAS ‘estimate’ statement. I quantiﬁed
the effect of oxygen on nos activity for each test isolate by plotting the natural log of N20
consumption rate against oxygen level. These slopes, which are equivalent to the
exponential decay constant, k, for untransformed data represent the sensitivity of nos
activity to oxygen. I subjected these data to analysis of covariance (isolate as main effect
and oxygen as covariate) using the GLM procedure in SAS and made a priori
comparisons using the ‘estimate’ statement (SAS Institute 1996).

Mean calculated oxygen concentrations in broth solution, after adding 0.25, 0.50,
and 0.75 mL atmosphere, were 0.053 1 0.007, 0.100 1 0.013, and 0.141 1 0.020
pmol ° L'l, respectively. These oxygen concentrations are well below the postulated 10
pmol ' L'l oxygen threshold for denitriﬁcation activity (Tiedje 1988). I found no evidence
for signiﬁcant oxygen consumption during these short term incubations, i.e. oxygen
concentrations, corrected for GC ﬂuctuations, did not vary during incubations. For

simplicity I report oxygen concentrations as mL of atmosphere added per vial.

51

RESULTS
Enumeration and isolation

I isolated a total of 156 denitriﬁers: 93 from the agricultural ﬁeld and 63 from the
successional ﬁeld. From the agricultural ﬁeld, I isolated 67, 21 and 5 denitriﬁers from
the 104, 10'5 and 10’6 dilution plates, respectively. From the successional ﬁeld, I isolated
58, 5, and 0 denitriﬁers from the 10“, 10-5 and 104’ dilution plates, respectively. There
were no differences in the number of viable heterotrophic colonies growing anaerobically
on R2A* plates inoculated with the two soils, in gas producers growing in R2A* broth,
nor in conﬁrmed denitriﬁers (Table 3.2). When results from all dilution series were
combined, a greater percentage of the heterotrophic colonies on plates from the

agricultural ﬁeld than from the successional ﬁeld were conﬁrmed as denitriﬁers.

Isolate characterization

Results of the cluster analysis are presented as a dendrograrn in Figure 3.1 and
taxa membership is provided in Table 3.3. Three broad groups are deﬁned at a Euclidean
distance of 0.53 units (A, B and C in Figure 3.1). The only reference strains in Group A
are from the alpha and beta subdivisions of the proteobacteria (Table 3.3). Group B, the
largest group, contains the majority of ﬁeld isolates but only four reference strains, two of
which are Bacillus species, the only Gram-positive reference strains. The other two
reference strains within Group B are of uncertain classiﬁcation (Tables 3.1 and 3.3).
Group C contains only 26 ﬁeld isolates, but many reference strains; all but three of these
reference strains belong to the gamma subdivision of the proteobacteria (Tables 3.1 and
3.3). One third (34 percent) of isolates from the agricultural ﬁeld, but only 14 percent
from the successional ﬁeld, clustered in Group A. Half (54 percent) of the isolates ﬁom

the agricultural ﬁeld and almost two-thirds (63 percent) of the isolates from the

 

52

Table 3.2. Mean number of anaerobic heterotrophic bacteria viable on R2A* agar, mean
number of gas-producers in R2A* broth, and mean number of conﬁrmed denitriﬁers for

soils from the conventionally-tilled agricultural ﬁeld and the never-tilled successional

 

 

ﬁeld at the KBS LTER site.
Agricultural Successional
Means, total number or percentage ﬁeld ﬁeld t test
Means
Anaerobic heterotrophs (x 10") 1.2 i 0.3 0.9 i 0.6 0.13 "5
Gas producers (x 105) 3.8 i 0.7 1.8 _1: 2.0 1.24 11.
Conﬁrmed denitriﬁers (x 105) 2.2 i 0.3 1.2 i 1.8 0.74 “8

Total number

Anaerobic heterotrophic colonies 410 400

Gas producers 157 84

Denitriﬁers 93 63
Percentage

Heterotrophic colonies that produced gas 38 21

Heterotrophic colonies that denitriﬁed 23 16

Gas-producers that denitriﬁed 59 75

 

Note: Values are means -_l~_ 1 SE (n = 3), total number of bacteria counted or isolated from
all plates with less than 300 colonies (sum of counts from 10“, 10's, and 10° dilution
plates) from three ﬁeld replicates, and percentages of these totals.

"3 P > 0.10.

53

Figure 3.1. Dendrogram, based on cellular fatty acid proﬁles, of 36 reference strains and
193 denitrifying bacteria isolated from soils from the conventionally-tilled agricultural
ﬁeld and the never-tilled successional ﬁeld at the KBS LTER site. Taxonomic groupings
at a coarse scale (0.53 Euclidean units) are indicated by the letters A, B, and C.
Taxonomic clusters at a ﬁner scale (33 species-level taxa) were deﬁned by grouping
isolates at the Euclidean distance represented by the dashed line (0.13 Euclidean units).
These clusters are identiﬁed by the numbers listed to the right. Isolates comprising each

taxa are listed in Table 3.3.

54

0.78 '
Euclidean distance 0.13 0

   

 

 

 

 

 

 

 

 

 

 

 

ill/viii“

 

i0"

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.1

 

 

 

 

 

 

 

 

 

55

Table 3.3. Denitrifying bacteria isolated from the conventionally-tilled agricultural ﬁeld and the never-
tilled successional ﬁeld at the KBS LTER site, and reference strains, grouped by taxa deﬁned by cluster
analysis (Figure 3.1).

 

 

 

 

Groups deﬁned
at Euclidean distance Soil isolates Reference
0.53 0.13 Agricultural ﬁeld Successional ﬁeld strains
A l 7, 9, 15, 25, 30, 36, 38, 175 Agra. tumefaciens (G41),

39, 40, 48, 49, 53, 55, 65,
66, 67, 72, 75, 79, 80, 83,
84, 91, 94,103, 104,111

Rhiz. sp. (OK-55), Ps. type 11
(G107, G143, G163, Gl88),
Achcycloclastes 21921

2 [2 [26,127, [28, 144,177,
188, 189, 192
3 35, 101
4 44
B 5 8 1_l_5,117,119,124, 125,
137,141,167, 187,193,
195
6 26, 37, 85,105 118,122,129,130,131,133,
136, 138, 139, 142, 158, 179,
190, 201, 202, 203
7 27, g_9_, 3_l, 32, 34, 43, Bacillus sp. (G192, G193)
56, 77, 86, 93, 112,
8 46
9 L2
10 89 169
11 52 184, 196
12 8; 120, I59
13 134, 135
14 22, 23, 24, 47, 50, 168 P.unknown type 3 (G108)
51, 54, 63, 73, 76, 81
15 140
16 88
17 88
18 _LQ
l9 Aquaspirillum itersonii
20 £113., [6,11,18,12, [63,172,114
2_0, A. 28.5.8.1 90. 69.
21 H
22 106, 107
C 23 33, 68 143, 160, 166, 180, P.flu. C 17400, 17561
181,182,183,186
24 161 P. ﬂu. 33512, P. ﬂu. F 17513
25 57, 64, 92 113, 121, 165 P.ﬂu. B 17467,
P.ﬂu. A 17552
26 P. aur. 17415, 17417,
P. ﬂu. 948, P. ﬂu. G 17386,
P. putida 17391, 39168
27 P. stutZeri JM300 (6 reps)
28 P. aeruginosa PAOl
29 B. pickettii PKOI,
P. putida 17472, 25571
30 70, 78, 90, 98, 99 191, 194 Alc.xyl.ssp. denit. (G65,G191),
Ale. faecalis (G148),
P.ﬂu. B 17812
31 114 P. chlororaphis 17810, 17812
32 Ale. xyl. NCIB 11015 (7 reps)
33 100

 

Note: Numbers following reference strains are ATCC accession numbers or other strain identiﬁcation
numbers as noted in Table 3.1. Isolates in bold were used as test isolates for nos assays. Isolates in italics
lost the ability to reduce nitrous oxide. Isolates underlined did not grow after storage in glycerol.

56

successional ﬁeld clustered into Group B. Group C contained 12 percent of the isolates
from the agricultural ﬁeld and 23 percent of the isolates from the successional ﬁeld.

Differences in community composition can also be seen at a ﬁner level of
taxonomic resolution. I used a Euclidean distance of 0.13 as the threshold for taxon
delineation since 1) all replicates of each control organism clustered at a distance < 0.13,
2) all reference strains (except P. putida and some P. ﬂuorescence biovars) clustered
within the same group at this distance, and 3) most reference isolates clustered only with
reference strains of the same or very similar species at this distance (Figure 3.1, Table
3.3). In addition, of the 14 taxa deﬁned at this distance that contained reference strains,
only four contained more than one “species” (taxa 1, 26, 29, 30).

Thirty-three taxa were identiﬁed at a Euclidean distance of 0.13 units. The 156
denitrifying isolates from the two ﬁelds represented 27 taxa (six of the 33 taxa contained
only reference strains; Table 3.3). Only eight of the 27 clusters containing ﬁeld isolates
also contained reference strains (Table 3.3). The 93 isolates from the agricultural ﬁeld
represented 22 different taxa; the 63 isolates from the successional ﬁeld represented 17
different taxa. Twelve taxa were common to both soils, but very few of these represented
similar proportions of their respective communities. The four numerically dominant taxa
in the agricultural ﬁeld (taxa 1, 7, 1‘4, and 20) were not common, if even present, in the
successional ﬁeld. For example, Bacillus was very common in the agricultural ﬁeld -- it
was isolated from the 104, 10'5 and 10'6 dilution plates -- but no isolates ﬁ'om the
successional ﬁeld clustered with Bacillus reference strains. Likewise, the four numerical
dominants in the successional ﬁeld (taxa 2, 5, 6, and 23) were isolated inﬁ'equently from
the agricultural ﬁeld. Taxa of intermediate dominance also differed in the frequency of
isolation between soils (Table 3.3). None of my isolates clustered with P. putida or P.

fluorescens bv. G, ﬁve of six non-denitrifer reference strains.

57

Denitriﬁer diversity, measured using the traditional Shannon-Weaver diversity

index (Shannon and Weaver 1949) was 2.39 for the agricultural ﬁeld and 2.34 for the

successional ﬁeld.

Nitrous oxide reductase activity

I used 31 isolates, representing 20 taxa, as test organisms (Table 3.3). Within taxa
containing more than one isolate I chose test isolates randomly. Under my test conditions,
all of these organisms produced large amounts of N20 in vials with 10 percent acetylene
in the headspace and negligible amounts of N20 in vials containing no acetylene (Figure
3.2). Since acetylene blocks nos activity, this combination of results suggests that nos is
active in these cultures. I conﬁrmed nos activity by observing rapid N20 consumption
following its addition to separate vials not containing acetylene (Figure 3.2A). The shape
of N20 production curves in the presence of acetylene was similar for all isolates (Figure
3.2A) but four (Figures 3.2B-D). These four isolates belonged to four different taxa that
clustered near each other within Group B (taxa 6, 7, 8, and 10; Table 3.3).

I was not able to measure the nos activity of seven taxa that contained ﬁeld
isolates because all isolates within these groups were either not viable after glycerol
storage or had lost the ability to reduce N20 (taxa 2, 4, 9, 16, 17, 18, and 21; Table 3.3), a
common phenomenon when denitriﬁers are repeatedly cultured (e. g. Gamble et al. 1977,
Tiedje 1994). Lack of isolate nos activity was indicated by identical N20 production in
vials with and without acetylene in the types of assays illustrated in Fig. 3.2 (data not
shown); i.e. isolates were not able to reduce N20.

When cultures shown to have active nos were transferred to vials containing ﬁ'esh

media, N20 consumption under anaerobic conditions was rapid after an initial lag phase

58

Figure 3.2. A. Typical patterns of nitrous oxide production by denitrifying isolates grown
in batch culture in the presence (closed circles) and absence (open symbols) of acetylene.
Nitrous oxide reductase (nos) activity is suggested by the combination of high N20
production in the presence of acetylene and no N20 production in the absence of
acetylene. Nos activity is conﬁrmed by rapid N20 consumption following injection of 70
pmol NZO'L'1 in the absence of acetylene (open squares). B-D. Atypical N20 production
patterns in the presence of acetylene for isolate numbers 46 (B), 77 and 85 (C), and 89
(D)-

59

 

 

 

0
0
B 1
0
6
oxhh
6e
m
MW...
0
2
..._....O
00000000
7654321
83 mom—Ea 533 + wow 5 0N2
e
mam. m
A WW. 1
am 0
0% 8
no
01 .0 de
..1 6.6
on n. mm
MT a
.. 0
ab 0 N2
2
_.. _ . 4 . .0
00000000
7654321
33 mmmmcacmﬁ3+ mm a _Onz

 

 

 

 

0
0
D 2
0
1m
0
1.91.
1.0
6
0
4
11. . . . . _ .0
0 0 0 0 0 0 0 0
7654321
63 833 963+ +8.3 5 Onz
0
.01
C
0
6
0
6
0
4
1.0
2
_d . _ . . _ 0
0 0 0 0 0 0 0 0
7 6 5 4 3 2 1
ionosonz

63 mom—Ea 2263+

Time (h)

Time (h)

Figure 3.2

60

(Figure 3.3). Nitrous oxide consumption rates, calculated as the slope of the linear portion
of the N20 consumption curves for each isolate, were variable, ranging from 2.85 to 14.4
pg NZO'h'l. One isolate (number 184) exhibited unusual behavior in that net N20
production was apparent prior to N20 consumption.

Once N20 consumption occurred, isolates were transferred to vials containing
four levels of oxygen. Oxygen inhibited nos activity of all test isolates (Table 3.4). For
25 of 31 isolates, the effect of oxygen was apparent with the addition of only 0.25 mL of
air (0.053 t 0.007 pmol oxygen ' L"). For four isolates (numbers 77, 89, 119, and 120),
the inﬂuence of oxygen only became evident at the higher oxygen concentrations (Table
3.4). Isolates number 46 and 184 showed no signiﬁcant differences in N20 consumption
rates with increasing oxygen according to AN COVA but a priori comparisons of N20
consumption rates at each oxygen level showed signiﬁcantly lower N20 consumption at
0.75 mL oxygen was added compared to the rate under anaerobic conditions (Table 3.4).

I measured exponential decay constants at two different points in time after N20
consumption began for seven of the isolates (numbers 33, 47, 65, 69, 85, 89, and 119) to
test whether variables that were likely to covary with time of sampling (e. g. population
size, nitrate concentration) affected the inﬂuence of oxygen on N20 consumption rate. In
all cases, while rates of N20 consumption at each oxygen level varied with time of
sampling (data not shown), k values were not different between sampling times (Table
3.5), i.e. the effect of oxygen on relative rates did not differ with sampling time. I used
only the data from the ﬁrst sampling time in analyses described below.

I also compared nos sensitivity to oxygen for isolates belonging to the same taxon
to test whether taxonomic divisions were physiologically meaningful. In all cases,
according to t tests conducted using the SAS ‘estimate’ statement following an analysis

of covariance, there were no differences in mean nos sensitivity to oxygen between any

61

Figure 3.3. Typical nitrous oxide consumption pattern for denitrifying isolates in
incubation vials to which 5 mL of 1.01 percent N20 was added. There was no net N20

production prior to rapid consumption.

62

I
100

I
80

 

60

40

 

 

_ _ _ _ 0
0 0 0 0 0
8 6 4 2

120 -
100 -

63 $83 .633 + mam :_ Omz

Time ’(h)

Figure 3.3

63

Table 3.4. The effect of oxygen on N20 consumption rates for 31 denitrifying isolates

representing 20 different taxa isolated from the conventionally-tilled agricultural ﬁeld and

the never-tilled successional ﬁeld at the KBS LTER site.

 

Taxa Isolate Field of mL air added

number number ori gina 0 0. 25 0.50 0. 75 F Wei
1 65 A 0.793 0.3 lb 0.200 0.180 78.91""
1 175 S 0.643 0.18b 0.140 0.11d 68.5"”
3 101 A 0.883 0.30b 0.140 0.090 134 *"'*
5 119 S 0.573 0.513b 0.40b0 0.290 10.8“”
6 85 A 0.203 0.1 lb 0.040 0.040 23.5“”
7 77 A 0.843 0.873 0.763 0.37b 12.5**"'
8 46 A 0.293 0.183b 0.193b 0.15b 2.35"s
10 89 A 0.253 0.223b 0.1 lb 0.1 lb 358*

l l 184 S 0.233 0.273 0.13b 0.07b 2.49m
12 120 S 0.393 0.293b 0.18b0 0.090 8.09'"
13 134 S 0.233 0.13b ND 0.1 lb 5.33*
14 47 A 0.503 0.28b 0.140 0.110 113 ***
14 168 S 0.583 0.3 lb 0.170 0.130 65.0“”
15 140 S 0.703 0.28b 0.19b0 0.060 l8.5"""*
20 69 A 0.163 0.10b 0.07b 0.05b 7.24‘"
22 106 A 0.503 0.20b 0.1 lb 0le 14.2““
23 33 A 1.373 0.88b 0.590 0.39d 78.0***
23 182 S 1.223 0.76b 0.440 0.420 6.66'"
24 161 S 1.443 0.76b 0.470 0.360 58.3“”
24 P. ﬂu. Fd R 0.983 0.70b 0.48b0 0.260 7.10””
25 64 A 0.963 0.84b 0.600 0.39d 49.01""
25 165 S 0.853 0.66b 0.440 0.32d 47.61"“
30 70 A 0:28a 0.16b 0.070 0.060 55.8“”
30 78 A 0.363 0.15b 0.12b 0.050 23.4***
30 90 A 0.493 0.19b 0.14b ND 27.1“”
30 98 A 0.333 0.1 lb 0.06b 0.05b 5.30*
30 99 A 0.573 0.27b 0.28b 0.140 40.8***
30 191 S 0.543 0.3 lb 0.23b 0.090 24.4“"
30 194 S 0.373 0.23b 0.120 0.06d 36.01""
31 114 S 1.303 0.99b 0.720 0.52d 99.7"”
33 100 A 0.733 0. 28b 0. 24b 0. 25b 148 ***

Nitrous oxide consumption rate°

(ng NZO'minute )

 

 

2A: agricultural ﬁeld, S— — successional ﬁeld, R= reference strain.

:Values within a row followed by the same letter do not differ at P < 0. 05.

°F statistics are for the interaction term time*oxygen 1n analyses of covariance (main
effect: oxygen, covariate: time); a signiﬁcant F statistic means the N20 consumption
rates for that isolate are not all equal. ns = not signiﬁcant, P > 0.05; * = P < 0.05; ** = P
d<0.;001 ***=P<0.0001.

°ATCC 1 75 13.

Table 3.5. Exponential decay constants, k, used to measure the sensitivity of nos to

oxygen for seven isolates measured at two different points in time aﬁer N20 consumption

began. These experiments test whether variables that are likely to covary with time of

sampling affect the inﬂuence of oxygen on N20 consumption rate. Results show that

sampling time did not inﬂuence k values.

 

Exponential decay constant, k

 

Isolate Sampling Sampling
number time 1 time 2 t test P
33 -1.77 i 0.25 -1.69 i 0.36 -0.20 0.89
47 -2.69 i 0.31 -2.03 i 0.29 -1.62 0.11
65 -1.98 i 0.25 -1.87 i 0.35 -0.28 0.78
69 -1.62 i 0.25 -1.69 i 0.23 0.21 0.83
85 -2.77 i 0.25 -2.68 i 0.44 -O.19 0.85
89 -1.50 i 0.24 -1.47 _+_ 0.29 -0.11 0.91
119 -0.97 + 0.31 -1.03 + 0.31 0.14 0.89

 

Note: t tests and their P values were calculated using the estimate statement in SAS

(n 2 2).

65

Table 3.6. Exponential decay constants, k, used to measure the sensitivity of nos to
oxygen for isolates belonging to taxa with two test isolates. Results show that there were

no differences in the k values between isolates belonging to the same taxa.

 

 

Taxon Isolate Exponential decay

number number constant. k tte§t P

l 65 -1.98 i 0.25

1 175 -1.93 j; 0.36 0.13 0.89
14 47 -2.69 j; 0.31

14 168 -2.67 i 0.24 0.05 0.96
23 33 -1.78 i 0.25

23 182 -1.69 i 0.26 0.23 0.82
24 161 -1.84 i 0.27

24 P. ﬂu. F ATCC 17513 -2.03 i 0.23 0.50 0.61
25 64 -1.23 i 0.25

25 165 -1.32 i 0.25 -0.26 0.80

 

Note: t tests and their P values were calculated using the estimate statement in SAS.

66

two isolates belonging to the same taxon (Table 3.6). Results from taxon 30 are
presented separately (Table 3.7) to account for the large number of pairwise comparisons
necessary to compare means among the seven test isolates.

AN COVA showed that there were signiﬁcant differences among the 31 test
isolates’ k values (Table 3.8), which ranged from 0.87 to 3.37 (Figure 3.4). As a rough
estimate of the denitn'ﬁer community k value for each soil, I multiplied the number of
isolates in each group by the mean k value for that group and then summed values for all
groups within each soil. The calculated k values for the denitrifying communities from
the two soils were not different: the mean k value for the denitrifying community from
the agricultural ﬁeld was 1.95 (95 percent conﬁdence interval = 0.12) and that for the

successional ﬁeld was 1.86 (95 percent conﬁdence interval = 0.20).

67

Table 3.7. Exponential decay constants, k, for the seven isolates belonging to taxon 30

and t test results to compare k values among the seven isolates. Results show that there

were no differences among k values for the seven isolates belonging to taxon 30.

 

 

 

 

 

 

Isolate Exponential decay
number constant. k t tests
Isolate number
70 78 90 98 99 191
70 -2.24 + 0.31
78 -2.36 + 0.25 0.30 (.77)
90 -2.56 + 0.49 0.46 (.65) -0.26(.79)
98 -2.16 + 0.26 ~0.18 (.86) 0.53 (.60) -0.61 (.55)
99 -1.75 + 0.35 -1.03 (.31) 1.39 (.17) -l.24 (.22) 0.94 (.35)
191 -2.11 + 0.31 0.28 (.78) 0.61 (.54) 0.67 (.50) -0.13 (.89) 0.76 (.45)
194 -2.42 + 0.25 -0.44 (.66) -0.16 (.87) -0.16 (.87) 0.69 (.49) 1.53 (.13) 0.76 (.45)

 

Note: Values in parentheses are P values associated with each 1 test; t tests and their P

values were calculated using the estimate statement in SAS.

68

Table 3.8. AN COVA table to determine the effect of isolate on the sensitivity of nos

activity to oxygen, k (main effect, isolate; covariate, oxygen).

 

 

Source of

mtion df MS F P
Isolate 30 2.00 31.72 0.0001
Oxygen 1 91.61 1451.98 0:0001

Isolate*oxygen 30 0.34 5.42 0.0001

69

Figure 3.4. Sensitivity of nos to oxygen for 31 denitrifying isolates representing 20 taxa
isolated from the conventionally-tilled agricultural ﬁeld and the never-tilled successional
ﬁeld at the KBS LTER site. The sensitivity of nos to oxygen was quantiﬁed by plotting
the natural log of N20 consumption rate against oxygen level. These slopes are
equivalent to the exponential decay constant, k, for the untransformed data that are
presented in Table 3.4. Error bars are i 1 SE. Taxa numbers below the x-axis are the
same as in Figure 3.1 and Tables 3.3 and 3.4. Numbers in parentheses above the x-axis
are the isolate number for those taxa with more than one test isolate. R* refers to the
reference strain P. ﬂuorescence F ATCC 17513. Isolate number is not included for those

taxa containing only one test isolate.

7O

 

 

 

 

 

 

 

 

 

 

lull. .. m
TIOIIL a m.
.IIOII. .. m
u b S.» - E
o u an: - .1
o 1 ea - om
TIIOII. 9m: .. 8
v o, . at-om
.r 0 l. .. mm
.1101 T NP
0 O a at r 8
ll 3% w on
1111 am: . on
loll. r :
11.11 «E u em
11 e:
.IIICII. ..
11.11 «we: r em
1+1. «9Q - mw
TIIIOII. as» . on
Tlolln. «we: . 8
11011 1 cm
rIII u or
I'll. .. mm
.1101 «we: - mm
III. 1 m
.IIOII. are - mm
Til. 1 M0
.1101 ..
1|.Ill. 1 m
T|.|||. 1 mp
5. a 5 m. 5 ._. s o
3 2 1 0

v: .5ng 9. mo: u_o >=>Ewcmm

Taxon

Figure 3.4

71

DISCUSSION

Denitriﬁer community composition was markedly different between soils at two
levels of taxonomic resolution. At a coarse scale, a greater proportion of isolates ﬁ'om
the agricultural ﬁeld than the successional ﬁeld clustered with reference strains belonging
to the alpha and beta subdivisions of the proteobacteria. Conversely, a greater proportion
of isolates from the successional ﬁeld than from the agricultural ﬁeld clustered with
reference strains belonging to the gamma subdivision of the proteobacteria (Table 3.3).
At a ﬁner scale of resolution, the distribution of isolates among 27 taxa also differed
signiﬁcantly between the two soils. Only 12 of these taxa were isolated from both soils.
The four dominants in each soil were not commonly isolated from the other soil and
isolates of intermediate dominance also differed in frequency of isolation between soils.

Denitriﬁer isolates from both the agricultural ﬁeld and the successional ﬁeld
differed in the degree to which oxygen inﬂuenced nos activities. My measure of nos
sensitivity to oxygen, the exponential decay constant for N20 consumption with
increasing oxygen level, k, varied from 0.87 to 3.37 (Figure 3.4). In other studies
showing differences in enzyme activity among denitriﬁer isolates, it was not always clear
whether differences were attributable strictly to differences in species physiology or
whether enzyme status differed among isolates at the beginning of experiments (Burth
and Ottow 1983, Abou Seada and Ottow 1985, Munch and Ottow 1986, Anderson and
Levine 1986, Munch 1989, Munch 1991). Denitriﬁcation enzyme status has since been
shown to be an important regulator of N20 production potential (e. g. Dendooven and
Anderson 1994). By ensuring that all denitriﬁcation enzymes were induced prior to
measuring nos sensitivity to oxygen, I were able to clearly demonstrate diversity in nos
sensitivity to oxygen among denitriﬁer isolates from two different habitats. In addition,

previous studies (including Carlson and Ingraham 1983, Ka et al. 1997) were conducted

72

on a small number of isolates and often using only isolates selected from culture
collections. I am not aware of other studies showing that the sensitivity of denitriﬁcation
enzyme activity to environmental regulators differs among members of natural denitriﬁer
communities.

Since nos activity is also regulated by nitrate, nitrite, and nitric oxide (e.g.
Hochstein and Tomlinson 1988), it is reasonable to consider whether there was any
interaction between oxygen and nitrogen substrate regulation in these incubations. I
addressed this issue by measuring the sensitivity of nos activity to oxygen at two different
points in time for a subset of isolates. If there had been an interaction between oxygen
level and nitrogen substrate concentration (or any other variable likely to covary with
time of sampling) 1 would expect a difference in the k values between the two curves.
Since there were no differences in k values for individual isolate incubations initiated at
two different points in time (Table 3.5), I conclude that oxygen is the dominant
regulatory control under my incubation conditions.

Despite differences in oxygen sensitivity of nos activity among isolates, I did not
ﬁnd any differences in nos activity regulation at the community level, using my weighted
average approach. There may, however, be some limitations to this means of estimating
community function. For example: 1) isolates may not represent all ﬁmctionally
signiﬁcant species in situ, 2) some of the 27 taxa containing ﬁeld isolates that lost nos
activity during culturing or lost viability following glycerol storage were not included in
community-level measurements, and 3) interactions among organisms, even if all
numerically and functionally dominant species are isolated, are not reproduced by
studying organisms in isolation. Some, or all, of these factors may have been important
in this study, especially in light of the differences I found in oxygen sensitivity of nos

between these same two communities using a soil enzyme assay, a method that does not

73

rely on culturing organisms (Chapter 2). Nonetheless, the physiological diversity among
isolates described here provides clear evidence for the community-level differences
described in Chapter 2; i.e. the results from Chapter 2 are dependent on there being
physiological diversity among denitriﬁer populations within these communities.

In addition, I found evidence for physiological diversity among these isolates that
may not be captured in the exponential decay constants. Four isolates exhibited
uncharacteristic N20 production curves in the presence of acetylene (Figures 3.2B-D).
Two isolates had unusual N20 consumption curves, suggesting that nos activity relative
to nor activity was delayed or inhibited in these isolates. Since in situ N20 ﬂux rates
depend on more than just the sensitivity of nos enzymes to oxygen, these other forms of ‘
physiological diversity may also inﬂuence in situ ﬂux rates.

Although it is becoming increasingly clear that isolates may not represent
numerically dominant bacteria in environmental samples (e. g. Torsvik et al. 1990a,
Richaume et a1. 1993, Tsuji et a1. 1995), there is no reason to doubt that the physiological
diversity present in this study would not also be found within any other subsample of
these two denitrifier communities. In fact, my isolate selection methods would, if
anything, tend to select for organisms with similar characteristics (e. g. Gorlach et a1.
1994), making my estimate of in situ physiological diversity conservative.

That a traditional measure of community diversity -- the Shannon-Weaver index -
- was not different for the two communities indicates an important distinction betweeen
diversity and community composition: the same level of diversity may represent
signiﬁcantly different community composition. This can be an important distinction with
implications for community function since community function may have more to do
with community composition than diversity per se (e. g. Grime 1997, Hooper and

Vitousek 1997, Huston 1997, but see Tilrnan et al. 1997).

74

CONCLUSION

My study indicates that microbial taxonomic diversity may have functional
signiﬁcance and supports ﬁndings from Chapter 2. Denitriﬁer community composition
was different in two geomorphically similar soils sampled from an agricultural ﬁeld and a
successional ﬁeld. Since culturing bacterial isolates under one set of conditions tends to
select for organisms with similar characteristics, the physiological diversity shown here
probably represents a conservative estimate of that which exists in situ. I also found
signiﬁcant diversity in the sensitivity of these isolates’ nos enzymes to oxygen. When I
calculated a mean, weighted measure of nos sensitivity to oxygen for each denitriﬁer
community, however, there was no difference between the two ﬁelds. The discrepancy ‘
between community-level measurements extrapolated from individual isolates versus
those measured for the entire community in a soil enzyme assay (Chapter 2) emphasizes
that differences in isolate physiology must be placed in the context of the entire
community in order to assess their importance to ecosystem-level ﬁmction. A complete
understanding of the effect of denitriﬁer community composition on N20 ﬂuxes will
depend on further study of the effect of environmental regulators, including those other
than oxygen, on the activity of nos and the other enzymes in the denitriﬁcation pathway.
My results, nonetheless, show that considering the effect of even a single regulator --
oxygen - on a single denitn'ﬁcation enzyme - nos - demonstrates that denitriﬁer
taxonomic diversity has clear potential to inﬂuence an important ecosystem function in

these soils.

PlantandSoil 170: 99-113. 1995.
© 1995 Kluwer Academic Publishers. Printed in the Netherlands.

Fatty acid methyl ester (FAME) proﬁles as measures of soil microbial
community structure

Michel A. Cavigelli‘, G. Philip Robertson2 and Michael J. Klug3

'W.K. Kellogg Biological Station and Center for Microbial Ecology, Michigan State University. Hickory Corners.
MI 49060. USA; 2116K Kellogg Biological Station and Department of Crop and Soil Sciences. Michigan State
University. Hickory Corners. MI 49060. USA and 3WK. Kellogg Biological Station and Department of
Microbiology and Public Health. Michigan State University. Hickory Corners. MI 49060, USA

Key words: fatty acid methyl ester (FAME) analysis. geostatisncs. principal components analysis. soil microbial
community

Abstract

Analysis of fatty acid methyl ester (FAME) proﬁles extracted from soils is a rapid and inexpensive procedure
that holds great promise in describing soil microbial community structure without traditional reliance on selective
culturing. which seems to severely underesdmate community diversity. Interpretation of FAME proﬁles from
environmental samples can be difﬁcult because many fatty acids are common to different microorganisms and
many fatty acids are extracted from each soil sample. We used principal components (PCA) and cluster analyses
to identify similarities and differences among soil microbial communities described using FAME proﬁles. We
also used PCA to identify particular FAMEs that characterized soil sample clusters. Fatty acids that are found
only or primarily in particular microbial taxa - marker fatty acids - were used in conjunction with these analyses.
We found that the majority of 162 soil samples taken from a conventionally-tilled corn ﬁeld had similar FAME
proﬁles but that about 20% of samples seemed to have relatively low. and that about 10% had relatively high.
bacterial:fungal ratios. Using semivariancc analysis we identiﬁed 21:0 iso as a new marker fatty acid. Concurrent
use of geostatistical and FAME analyses may be a powerful means of revealing other porential marker FAMEs.
When microbial communities from the same samples were cultured on R2A agar and their FAME proﬁles analyzed.
there were many differences between FAME proﬁles of soil and plated communities. indicating that proﬁles of
FAMEs named from soil reveal portions of the microbial community not culturable on RZA. When subjected to
PCA. however, a small number of plated communities were found to be distincr due to some of the same proﬁle
characteristics (high in 12:0 iso, 15:0 and 17:1 ante A) that identiﬁed soil community FAME proﬁles as distinct.
Semivariance analysis indicated that spatial distributions of soil microbial populations are maintained in a portion
of the microbial community that is selected on laboratory media. These similarities between whole soil and plated
community FAME proﬁles suggest that plated communities are not solely the result of selection by the growth
medium, but reﬂect the distribution. in situ. of the dominant. culturable soil microbial populations.

Introduction

Recent evidence suggests that culturable soil microor-
ganisms may represent a tiny. possibly ecologically
unimportant portion of the overall diversity present
in most soils. DNA reassociation kinetics. for exam-
ple. suggested the presence of at least 4,000 bacterial
strains in 30g of a forest soil from southern Norway.
with culuirable strains representing less than 1% of the
t0tal present (T orsvik et al.. 1990 ab). Such studies are

consistent with the typical ﬁnding that about 100 times
more soil bacten'aareseenwithdirectmicroscopythan
are found using classical plating methods. Results such
as these emphasize the need for new techniques that
allow us to describe soil microbial community struc-
ture without the usual reliance on selective culturing.
an approach that appears to severely underesumate
the actual diversity of soil microbial communities.
In recent years a number of useful approaches have
been developed to address this challenge. including

75

100

16S rRNA probes. RFLP analyses of the products of
PCR primers. %G + C proﬁles. and fatty acid methyl
ester (FAME) proﬁle analysis.

In this paper we present the results of an in situ
assessment of the Spatial distributions of soil micro-
bial communities - part of an effort to better understand
patterns. causes. and consequences of microbial diver-
sity in soils of agronomic signiﬁcance. We used FAME
proﬁle analysis for this task because of its unique abil-
ity to characterize whole communities rapidly and at
relatively low cost.

The value of FAME analysis arises from the facts
that there are a great number of different kinds of fatty
acids in the lipids of microorganisms and that different
organisms have different combinations of these fatty
acids. Because fatty acids can be readily volatilized
following methylation. they can be readily analyzed
by gas chromatography (e.g. Moss et al.. 1980; Moss
1981; Vestal and White. 1989). Many microbial iso-
lates or taxa have unique FAME proﬁles (e. g. Mayber-
ry etal.. 1982; thjoen ct al.. 1986). and the technique
has been adapted to the examination of mixed commu-
nities from both sediments (Bobbie and White. 1980;
Findlay and White, 1983; Hogg. 1984: Perry et al..
1979 and Volltman et al.. 1980) and soils (pers. comm.
H. Garchow and M. Klug. Michigan State University;
Zelles et al.. 1992: Zelles and Bai. 1993).

The interpretation of proﬁles from whole soil com-
munities can be difﬁcult because many fatty acids
are common to different microorganisms and because
there are hundreds of different fatty acids in environ-
mental samples. especially in agricultural soils (Zelles
et al.. 1992). Thus far. FAME proﬁle analysis has been
largely limited to qualitative and univariate descrip-
tions of the fatty acids present in environmental sam-
ples. We show in this paper that multivariate and geo-
statistical approaches to proﬁle analysis can substan-
tially aid the interpretation of FAME proﬁle data. We
also demonstrate the extent to which plated commu-
nities are representative of whole soil microbial com-
munities by comparing results ﬁ'om whole soil FAME
proﬁles with those of plated microbial communities
extracted from the same soils.

Material and methods
Study site

This study was conducted at the WK. Kellogg Bio-
logical Station's Long-Term Ecological Research Site

76

in Row-Crop Agriculture. located in SW Michigan. in
the northern portion of the US. corn belt. Soils of the
site are Typic Hapludalfs (Kalamazoo series). sandy
loams of moderate to high fertility. The Ap horizon
has a mean CEC of 5.7 meq 100g". a mean pH of
7.0 (1:2 water), and a mean total C content of 1.0%
(unpublished data).

Soil sampling

We collected 54 soil samples in July 1991 from an
80m transect in each of three replicate 1 ha agronomic
plots (162 total samples) planted to conventionally-
tilled corn (Zea mays L). Transects were placed 10cm
from and parallel to a chosen row of corn. Soils were
sampled to 10cm depth at locations along each tran-
sect following a nested sampling scheme designed to
minimize the number of samples required to capture
statistically signiﬁcant patterns of spatial autocorrela-
tion; sample intervals were as small as 2 cm. Sampled:
soil was stored in plastic bags in a cooler immediately
upon sampling and then later sieved (2 mm). mixed.
and subsampled within 12 h for the analyses described
below.

FAME analysis

Subsamples for FAME analyses (5 g) were stored at
4°C in ashed test tubes for 24 h and then processed
according to the Microbial Identiﬁcation System (MIS;
Microbial ID Inc. 1992) standard protocol. First. lipids
were saponiﬁed by adding 5.0 ml. 3.25 M NaOl-l in
methanol to each soil sample. then mixing. heating in
a 100° water bath for 5 trtin. mixing again. and heating
in the water bath for an additional 25 min. The sam-
ples were then methylated by adding 10 mL of 3.25 N
HCl in methanol. mixed. placed in an 80°C water bath
for 10 min, and then rapidly cooled in ice water. The
FAMEs were extracted from this solution by adding
1.5 ml. of one part methyl ten-butyl ether in one part
(v:v) hexane and placing the closed tubes on a rotary
mixer for 10 min. The top. organic phase was trans-
ferred by pipet to an ashed test tube and then washed
using 3.0 mL of dilute NaOH. The organic phase was
then transferred to a GC vial for subsequent analysis by
gas-liquid chromatography using an HP 5890 (Hewlett
Packard. Rolling Meadows. IL. USA) equipped with
an HP Ultra 2 capillary column (crosslinlted 5% Ph Me
silicone. 25 m x 0.2 mm x 0.33 mm ﬁlm thickness)
and a flame ionization detector. One person can pre-
pare 40 samples for analysis by gas chromatography
in under 4 h using this protocol.

Table 1. Market fatty acids. From Envm (1973). White (1983).
Harwood and Russell (1984). Jantzen and Bryn (1985). and Vestal
and White (19891

 

Eubacteria

14.03011 15:01: 17111! 17:0cyc 18:1cisll
19:0cyc 16:0br10 15:1at6 15:1at8 17:1at6
17:1at8 15zliso3 15:11:07

u'ansmonounsaturatedbranchedandsu'aigl‘u l6and 17C
(eubacteriado not. in general. contain polyunsaturated fatty acids)

Gram negative eubacteria
01-1 fatty acids (usually 3 011)

Gram positive eubacten'a (but also found in Grant negatives)
branched fatty acids (iso. anteiso)

Eultaryotes
12:0 16:1 at7 18:2cis9.cis 12 18:1 at9
0-1823 cis 9. cis 12. cis 15 polyunsaturated fatty acids with > 20 C

Protozoa
20:3 at 6 20:4 at 6

Microfauna
y- 1 8: 3

All organisms

14:0 16:0 18:0

 

Since whole cell fatty acids have proven sufﬁcient
to distinguish microbial communities from soils which
differ only in the type of agricultural management
to which they have been subjected (Klug and Tied-
je, 1993). we did not separate the phospholipids from
the rest of the lipid fraction prior to saponifying the
lipids.

We report our results (e.g. Table 1) using standard
FAME nomenclature: the number of C atoms in the
fatty acid is indicated by the number before the colon
and the number after the colon indicates the number of
double bonds. “Ante" means anteiso. and “br” means
branched. either at the iso or anteiso positions. “Cyc”
refers to the cyclo propane analogue. and numbers fol-
lowing a C indicate the location of the epoxy bond.
Use of “at” indicates that cis or u'ans conﬁguration is
not known and the number(s) following the cis. trans.
or at designation indicate(s) the position of the double
bond(s) relative to the carboxyl end of the molecule.
The number before an OH refers to the location of a

77

101

hydroxy substitution relative to the carboxyl end of the
molecule. A capital letter at the end of a monounsatu-
rated acid indicates that the position of the double bond
is not known but that the bond is at a different loca-
tion than for other monounsaturated FAs of the same
chain length and branching pattern that have a differ-
ent letter designation. Sif means “sum in feature” and
indicates that more than one FAME has a particular
retention time. For example. sif 13 is a combination. in
unknown proportions. of 19:1 trans 11. an unknown.
and 19:0 cyc C9- 10. Interpretation of proﬁles has been
aided by the use of FA markers - those FAs found only
or principally in particular groups of organisms (Vestal
and White. 1989); marker FAs identified by various
investigators are listed in Table 1.

Plate cultures

Subsamples used to culture microbial communities
(1g) were refrigerated at 4°C for up to 24h prior to '
preparing a dilution series in phosphate buﬁer (pH 7.2).
One ml. from each 10'2 dilution tube was transferred
to a 150 x 10 mm plate containing R2A agar. Plates
were incubated at room temperature and incubated for
36—40h; growth was rapid. Plates were then stored at
4°C for no more than 48h. The entire plated commu-
nity was scraped into an ashed test tube and returned
to the refrigerator for up to 48h. FAME analyses were
conducted as for soil samples but using smaller quan-
tities of reagents.

Microbial biomass

From the same dilution series as used for plating com-
munities. acridine orange direct counts (AODC) were
prepared for each sample from one of the transects.
Acridine orange (0.5 mL of 1% solution) was added
to 10 ml. of the 10’3 dilution. Three ml. of this solu-
tion were ﬁltered through a 0.2pm Millipore mem-
brane and the membrane was then washed with 6 mL
of isopropanol. The ﬁlter was placed on a slide and
20 ﬁelds were counted at 63X. Counts were converted
to bacterial biomass values based on an average size
of 0.196 mm3 bacterium". calculated from measure-
ments made previously on bacteria from the same site
(pers. comm. D. Harris. Michigan State University).

Statistical analysis

We used principal components analysis (PCA: SAS
Institute. 1991) to compare FAME proﬁles among soil
samples and, separately. among plated communities.
PCA. like other multivariate statistical methods. is

102

used to summarize data in which multiple variables
have been measured for each sample. In PCA the orig-
inal variables (FAMEs in this case) are orthogonally
transformed into a new set of uncorrelated variables
called principal components (PCs). Since each PC is a
linear combination of the original variables. all origi-
nal variables are represented in each PC. The degree to
which each PC is inﬂuenced by each original variable
is its eigenvector loading, the sign of which indicates
the manner in which the original variable influences the
PC. The variance associated with each PC decreases
in order. so that much of the variability of the origi-
nal variates may be accounted for in the ﬁrst few PCs.
A subset composed of those variables most important
in discriminating among samples (i.e. with the high-
est loadings in the primary PCs) can be selected and
used for further analysis and interpretation (Digby and
Kempton. 1987: loliffe. 1986 and Liu and Keister).

We used the correlation matrix rather than the
covariance matrix in PCA since the standard deviations
of some of the individual FAMEs were large compared
to the means (Joliﬁ'e. 1986). Results of PCA may be
presented as biplots - plots of eigenvector loadings of
each variable. usually scaled and superimposed on a
PC plot of the sample points. The farther a variable
plots from the origin. the more inﬂuential it is in a par-
ticular PC. The relationship among variables can be
determined by directional cosines of the angle formed
between lines drawn from the origin to each variable.
Variables at 90° to each other have no effect on each
other (cos 90° = 0) and variables at 180° to each other
have opposing effects (cos 180° = 1).

We also subjected the data to hierarchical clus-
ter analysis using NTSYS (Applied Biostatistics Inc.,
1992). We constructed dendrograms for most matri-
ces for which we used PCA. using all combinations
of Euclidean and average taxonomic distance metrics.
and UPGMA (unweighted pair-groups method using
arithmetic averages) and single linkage. We chose link-
age methods that have been shown to be biased to the
formation of different types of clusters (Everitt, 1980;
Milligan and Cooper. 1987). Cluster analysis has the
advantage over PCA in that all dimensions of a data
matrix can be represented in a two dimensional dendro-
gram. but the disadvantage that variables contributing
to the clustering cannot be identiﬁed (Everitt. 1980;
Milligan and Cooper. 1987).

During processing and FAME analysis of samples.
seven whole soil and 19 plated community subsamples
were lost or did not provide sufﬁcient material for anal-
ysis; multivariate statistical analyses were conducted

78

on the remaining 136 samples shared by both soil and
plated community data sets.

We used semivariancc analysis. a geostatistical
technique, to describe the degree of spatial autocorre-
lation of individual FAMEs. i.e. to quantify the degree
to which soil samples taken close to one another are
more similar with respect to quantities of individual
FAMEs titan samples taken farther apart. Such infor-
mation can help to identify the scale at which controls
on microbial populations containing particular FAMEs
operate (Robertson and Gross. 1994: Trangmar et al..
1985) as well as to identify FAMEs unique to speciﬁc
microbial communities.

Semivariance. y(h). is deﬁned mathematically as

l N(h)
var) = EMT) é; [2(a) —- 2(2. + Ml2

where z(xi) is the value of a variable for a sample taken .
at point x. and 20:; + h) is the value of the variable for
another sample taken at a distance. or lag interval. h
from point x.. N(h) is the total number of sample pairs
separated by a distance h (Webster. 1985).

Semivariance analysis results in a variogram (e.g.
Fig. 4) of semivariancc for all possible lag intervals
within a given spatial domain. If a continuous variable
is sampled at an appropriate scale. 7(h) decreases as h
decreases to zero; a variate is perfectly autocorrelated
with itself at the same location. The population vari-
ance is estimated by the sill, that portion of the curve
where 7(h) is constant. The y-intercept is called the
nugget variance. The portion of the population vari-
ance due to spatial su'ucttue is the difference between
the sill and the nugget variance. The lag distance at
which the sill is reached is the range and indicates the
distance over which a variate is autocorrelated (Web-
ster. 1985).

Results and discussion

We present FAME data in units of percent of total
FAMEs within a sample. The MIS protocol we adapted
to whole soil samples is highly reproducible; e.g. the
standard deviations for ﬁve subsamples taken from a
well-mixed sample from the same corn ﬁeld were very
low compared to the means (Fig. 1).

On average. 24 different FAMEs were detected in
each whole soil sample. Fifty-six FAMEs were identi-
ﬁed in total (Table 2) and more than 80% of these were
detected in the ﬁrst 12 samples. indicating that most of

79

103

 

PereentofTotal

 

sif 3

16:0 180
16:1 C18 7
16:1 C18 9

15:0
16:1 C15 11

100

11:0 ISO
10.0 30H
120 ISO
12:0

13:0 ISO
13:0 ANTE
12:0 20H
12:0 30H
14:0180
14:0

15:0 180
15:0 ANTE

16:0

15:0 ISO 3011

 

21:0'

18:0
22:1 C15 17

17:0 180 30H

sif 9

18:1 C18 9
511 10

511' 13

17:0 ISO
19:0 CYCLO C11-12

17:0 ANT E
721 C18 9
7:0 CYC
6:0 2011
6:0 30H
86.12.14
18:0 2011
18:0 30H

20:1 C15 17

17:1 ISO 0

18:3 C

Fig. I. Musdstandarddeviations(WbyertorbarstFAMEsforﬁvesubsamplesﬁornonehornogeniaedsampleofsoiltaken

ﬁomacornﬁeld.

the diversity detected in this study was present within
averysmallportionofthesmdyarea.

Zelles et a1. (1992) found that phospholipid
palmitic acid. 16:0. was highly correlated with a num-
ber of measures of microbial biomass. We did not
ﬁnd any relationship between AODC-derived biomass
measurements and the proportion of 16:0 or any other
FAME common to all organisms or any FAME speciﬁc
to bacteria.

Principal components analysis

FAME proﬁles of whole soil communities included
FAMEs with GC retention times of up to 23 min;
those of plated communities included only FAMEs
with retention times of less titan 17 min. We con-
ducted PCA separately on two whole soil communi-
ty data matrices: a complete matrix including all 56
FAMEs and an abbreviated matrix that included only
those 46 FAMEs with retention times less than 17 min.
Because the results of the PCA were very similar for
both the complete and abbreviated soil proﬁles. we
present results only from the abbreviated proﬁles here.
Since a number of FAMEs speciﬁc to plants (Vestal
and White, 1989; White. 1983) have retention times
ontheGCgreaterthanZBminusingthisprotocol,

the lack of difference between complete and abbrevi-
ated proﬁles is evidence that FAMEs from roots and
other plant materials were probably not an important
component of the FAME proﬁles in these samples.

PCA of FAME proﬁles from whole soil communi-
ties showed that there were three distinct clusters of
soil samples (Fig. 2A). The symbols used in Figure 2A
and other PCA plots (Figs. ZB-C. 5A-B) serve only
to identify soil samples that clustered together; since
the same symbol is used for each sample throughout
this paper comparisons can be made among plots. All
but one of the 25 samples forming the small cluster
of stars in Figure 2A were taken within a 58 m sec-
tion of one of the three transects. indicating a patchy
distribution of microbial conununities at a large scale.
The soil samples represented by diamonds were more
evenly disuibuted across all three transects. indicating
spatial variability on a different scale.

Since only 39% of the total variability was
explained in the ﬁrst three PC dimensions (Table 3).
individual soil samples may be less similar than they
appear in Figure 2A. Nonetheless. the separation of
samples into three distinct clusters can be interpreted
by identifying the individual FAMEs with the highest
loadings in the eigenvectors of the PCs along which
clusters are separated. This type of analysis is usually

104

80

TableZ. Fattyacidsfoundinsoilandinplatedcommunities.1'hosefattyacidsfoundinboth
communitiesareindicasedinboldface.Fattyacidswithretennontimes>17minsontheGC
columnandtemovedforthewholesoilanalyseeuelistcdasasepamegroup

 

Monounsaturated cis Saturated Even Saturated Odd 3 and 12 OH
16:1 cis 7 101) 9:0 10:0 3 08
16:1 ch 9 12:0 15:0 12:0 3 0!!
16:1 cis 11 14:0 17:0 15:0 iso 3 OH
17:1 cis 10 16:0 16:0 3 OH
18:1 ch 9 18:0 Branched: anteiso 17:0 iso 3 OH
18:1 cis 15 13:0 ante 18:0 3 OH

Branched: iso 15:0 lite 18:0 12 OH

Monounsaturated ND 11:0 in 17:0 ante

(conﬁguration 12:0 to 19:0 ante Retention ﬁne
not determined) 13:0 ho > 17 mins
16:1 ho G 14:0 to Polyunsaturated 20:1 cis 17
17:1isoG [5:080 18:3 cis6. 12. 14 20:1at9
17:1 ate A 16:0 to 20:4 cis 14 21:0 iso
16:1 2 OH 17:0 to 20:5 cis 17 2M)
allO(18:1cisll. 19:0bo 21:1cisl8
18:1trans9. 20H 21:1 cisl7
18:1 trans 6) Cyclo 16:0 2 OH 22:0 3 OH
17:0eyc 18:020H 24:1 at9
25:0 2 OH
Plated communities
Monounsaturated ND Monounsaturated cis Branched: iso Cyclo
(conﬁguration 16:1 cis 9 11:0 to 17:0 eye
notdetennined) 18:1 c139 12:0 in 19:0cch11o12
16:1 iso E 13:0 ho
16:1 in G Monounsaturated trans 14:0 ho 2 0H
16:1 A 19:1 trans 7 15:0 in 12:0 2 OH
17:1 iso 8 16:0 to
17:1 iso H Saturated Even 17:0 be 3 0H
17:1 ate A 12:0 19:0 in 10:0 3 OH
17:1 B 14:0 12:0 3 OH
sif 5 (17:1 iso 1. 16:0 Branched: anteiso 17:0 in 3 0H
17:1 ante 8) 18:0 13:0 ante 18:0 3 OH
sﬂ'l(l8:l cis 11. 20:0 15:0ante
18:1 trans 9. 17:0 ante
18:1 trans 6) Saturated Odd 19:0 lite
15:0
17:0

 

presented as a biplot. but the large number of vari-
ables in this data set precluded clear presentation of the
biplot. so we tabulated those variables and their load-
ings that had the greatest inﬂuence (loadings of PC 1 >
|0.20|) on the separation of samples into the two distinct
clusters of circles and stars (Table 4). The principle dif-
ferences between samples forming the small cluster of

stars and the rest of the samples is that samples rep-
resented by stars had a greater proportion of 12:0 iso.
15:0. 16:1 iso G, 17:1 ante A. and 16:12 OH and a
smaller proportion of 12:0. 14:0. 15:01so. 16:0, and
18:0 2 OH than did the rest of the samples. No of the
monounsaturated FAMEs more common in samples
represented by stars (16:1 iso G and 17:1 ante A) are

81

105

Table 3. Proportion of variance accounted for by eigenvalues of the correlation matrices with
values 2 1 for principal component analyses

 

Proportion of Cumulative Proportion of
Variance Explained Variance Explained
Eigenvalue by Eigenvalue (%) by Eigenvalues (%)

 

Whole Soil: 9.14 19.9 19.9
full data 5.11 11.1 31.0
matrix 3.59 7.8 38.8
(Fig. 2A) 2.16 4.7 435
2.03 4.4 47.9
1.80 3.9 51.8
1.62 3.5 55.3
1.55 3.4 58.7
1.47 3.2 61.9
1.33 2.9 64.8
1.24 2.7 67.4
1.19 2.5 70.0
1.05 2.3 72.3
1.02 2.2 745
Whole Soil: 14 6.57 46.9 46.9
selected FAMEs 2.68 19.2 66.1
(Fig. 28) 2.36 16.8 82.9
Whole Soil: 3.69 36.9 36.9
grouped FAMEs 1.72 17.2 54.0
(Fig. 2C) 1.27 12.7 66.7
Plated Communities: 12.5 30.5 30.5
full data matrix 5.14 12.5 43.0
(Fig. 5A) 3.07 7.5 50.5
2.73 6.7 57.2
2.55 6.2 63.4
1.88 4.6 68.0
1.64 4.0 72.0
1.60 3.9 75.9
1.37 3.3 79.2
1.17 2.9 82.0
1.09 2.7 84.7
Plated Communities: 4.37 39.8 39.8
11 selected FAMEs 3.31 30.1 69.8
(Fig. 58) 2.03 18.4 88.3
Fined Communities: 4.57 45.7 45.7
grouped FAMEs 1.74 17.4 63.1

1.13 11.3 74.4

 

106

 

 

 

PC3 4

82

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IO'ﬂrﬂCOG)

h.—

0

Monounsaturated cis
Monounsaturated ND
Saturated Even
Saturated Odd

iso

ante

Cyclo
Polyunsaturated

2 0H

3 OH

Soil Samples

Soil Samples

Soil Samples

 

 

 

1011610061)

zit-Xv..—

0

10:0

12:0 iso

12:0

12:0 30H
14:0

15:0 iso

15:0 ante
15:0

16:1 cis 9
16:0

17:1 amtc A
16:0 20H
16:0 30H
sil 13

Soil Samples
Soil Samples

Soil Samples

 

 

Fig. 2. PCA plot and biplots for soil community FAME proﬁles. A. PCA plot for whole soil community FAME proﬁles containing 46 FAMEs.
B. Biplot for whole soil community FAME proﬁles containing 14 FAMEs selected using PCA C. Biplot for whole community FAME proﬁles

containing 10 groups ofFAMEs. Parametetsare listed in Table 3.

derived from FAs that are markers for bacteria (Perry
et al.. 1979). and one saturated FAME less common in
these samples ( 12:0) is a marker for eukaryotes (White.
1983). Since fungi have much higher biomass titan any
other eukaryotes in agricultural soils (e.g. Brussaard et
al.. 1990). the soil samples found in the small cluster

of stars probably have greater bacterial:fungal biomass
ratios than do the majority of samples. Samples denot-
ed by diamonds. which fall on the other end of the PC
1 axis. are. by similar reasoning, likely to have rela-
tively low bacterialzfungal biomass ratios. In addition.
our ﬁndings support Jantzen and Bryn’s (1985) sug-

83

107

Table 4. FAME proﬁle characteristics most inﬂuential in distinguishing the soil samples
represented by stars from those represented by open circles in Figures 2A (whole soil
communities) and 5A (plated communities). Those FAMEs that distinguished soil samples
from both soil and plated communities are indicated in bold. All FAMEs which had loadings
of more than |0.20| (an arbitrary threshold) in the relevant eigenvectors are included in this
table. ‘lhe value of the loadings are indicated in columns following each FAME

 

Stars high in Loading Open circles high in Loading
Whole 12:0 to [Gram pos?) 0.29 12:0 [eukaryotes] -0.20
Soil 15:0 0.30 14:0 [all organisms] -0.21
16:1 iso G [eubacteria] 0.27 15:0 iso -0.23
17:! ante A [eubacteria] 0.29 16:0 [all organisms] -O.29
16:1 2 OH 0.24 18:0 2 OH -0.21
Outliers high in Majority high in
Plated 12:0 iso [Gram pos'?] 0.40 10:0 3 0H [Gram neg] -O.20
15:0 0.40 16:1 cis 9 -0.24
17:1 ante A [eubacteria] . 0.40
19:0 iso [Gram pos?] 0.41

 

gestion that branched fatty acids are not good markers
for Gram positive bacteria since soil samples had both
high and low proportions of some of these putative
Gram positive markers (Table 4).

Since the amount of variance explained by the ﬁrst
three dimensions of the PCA was not very high, we ran
PCA on matrices with fewer variables. We reduced the
number of variables using two different approaches.
First. we took advantage of the potential offered by
PCA to eliminate variables that do not signiﬁcantly
explain total variance. We used Kaiser’s rule (Joliffe.
1986) - that only as many variables as there are eigen-
values 21 need to be kept for further analysis when the
correlation matrix is used in PCA - to determine that
14 of the FAMEs should be retained. We chose to keep
those 14 FAMEs with loadings > |0.20| in both PC 1
and PC 2 and those FAMEs with the largest loadings
(absolute value) in either of the ﬁrst two PC dimen«
sions. Results and the FAMEs used in this analysis are
presented in Figure 28. Because of the greatly reduced
dimensionality. presentation of results as biplots was
possible. but only two dimensions are illustrated since
three dimensional biplots did not show additional clus-
tering of sample points, and two dimensional biplots
were easier to read. In Figure 28. samples identiﬁed
as stars again formed a distinct cluster as did samples
denoted by diamonds. and more than twice the matrix
variability explained in Figure 2A was explained by
reducing the dimensionality (Table 3). PC 1 is a con-
trast of 12:0 iso, 15:0, and 17:1 ante A vs. 12:0, 14:0.

15:0 iso. 15:0 ante, 16:1cis 9, and 16:0. Since both 12:0
(White, 1983) and 16:1 cis 9 (Erwin. 1973) are mark-
ers for eukaryotes. samples represented by stars. again.
seem to have relatively high bacterialzfungal biomass
ratios and samples denoted by diamonds seem to have
relatively low bacterialzfungal biomass ratios. Thus.
differences in bacterialzfungal biomass ratios among
samples were highlighted by eliminating less inﬂuen-
tial variables.

We also conducted PCA on a reduced dimension
data matrix composed of FAMEs grouped according to
similarities in their structure (Fig. 2C). These group-
ings. similar to those used by Zelles et al. (1992),
reﬂect differences in the metabolic pathways required
to produce FAs and, as such, may be taxonomically
useful. We summed all FAMEs belonging to a particu-
lar group (Table 2) for each soil sample. The proportion
of matrix variance explained was now 54 and 67% in
two and three dimensions. respectively. Results were
very similar to those for ungrouped data - samples
represented by stars were distinctly different from the
other samples. open circles formed a gradient along PC
2, and the diamonds were still grouped, albeit loosely
- indicating that the groups into which we categorized
FAMEs may be biologically meaningful.

Interpretation of the PCA of grouped FAMEs is
consistent with that for ungrouped FAMEs. In the
ungrouped data (Table 4). samples represented by
stars harbored greater proportions than did the major-
ity of samples of three monounsaturated FAMEs with

108

Euclidean 0 A u._—£Dm ,8

 

 

 

 

 

 

 

 

L< <

Fig. 3. Dendrograms derived from cluster analyses for whole soil
communities. A. Euclidean clustering. UPGMA linkage. B. Average
taxonomic distance clustering. single linkage. The same symbols
used in Figures 2 and 5 are used for each soil sample.

unknown conﬁguration (16:1 iso G. 17:] ante A. 16:1
2 OH) and a saturated FAME with an odd number of
C atoms (15:0). Samples represented by stars also had
lower amounts of a FAME with a hydroxyl substitu-
tion at C number 2 (18:0 2 OH). These differences
are reﬂected in Figure 2C in that monounsaturated
ND and saturated odd FAMEs are contrasted with 2
OH FAMEs. Other differences among samples list-
ed in Table 4 were not found when grouped FAMEs
were subjected to PCA: this is because not all FAMEs
belonging to a particular group were distributed equal-
ly among soil samples so that. for example, samples
that were high in 12:0 iso were also low in another iso
FAME. 15:0 iso.

Cluster analysis

Results from only two of eight cluster analyses con-
ducted on soil community FAME proﬁles are present-
ed (Fig. 3) since all dendrograms were essentially
identical. Consistent results such as these, in light of
the fact that UPGMA and single linkage methods are
biased toward the creation of different types of clusters
(Everitt. 1980; Milligan and Cooper. 1987), suggest
that sample groupings identiﬁed by cluster analysis
reﬂect real similarities and differences among FAME
proﬁles of these soil samples. All eight cluster analyses
also reﬂected cluster patterns found in PCA plots. Soil
samples identiﬁed by stars and by open circles in PCA
formed separate clusters in all dendrograms and sam-
ples identiﬁed as diamonds formed separate clusters
only in the cluster analyses of the ungrouped data sets.
Cluster analyses therefore corroborated the clustering
identiﬁed in the ﬁrst two and three PC dimensions for ‘
the PCA of ungrouped and grouped data sets. Unlike
PCA. however, cluster analysis does not provide a
direct means of identifying variables that contribute
to the separation of samples into different clusters.

Geostatistical analyses

Our sampling scheme allowed us to capture previously
unrecorded small-scale autocorrelation in some com-
ponents of the numerically dominant portions of soil
microbial communities (Fig. 4A). All eight FAMEs
that exhibited spatial autocorrelation at the scale mea-
sured (minimum lag 0.07 m) exhibited a range < 1.00
m and most often < 0.20m (Table 5). This suggests that
controls on the distributions of organisms with these
FAMEs are also acting at these spatial scales (Trang-
mar et al.. 1985). For example, the distribution of
population(s) for which 15:0 ante is a marker appears
to be controlled by some factor or set of factors that
are acting at a scale of about 0.1m (Table 5, Fig. 4A).
Two FAMEs that are known to be markers, 15:0 ante
(eubacterial marker). and 16:0 2 OH (Gram negative
marker) showed a signiﬁcant pattern of spatial autocor-
relation (Fig. 4A, Table 5). FAMEs that had a unique
range of autocorrelation are probably also markers - at
least among samples used to construct the variogram
- since a unique range of autocorrelation identiﬁes. by
deﬁnition, a variable with a unique spatial distribution.
This appears to be the case for 21:0 iso which was the
only FAME with a range similar to 0.35m (Table 5).
The population(s) for which a FAME identiﬁed as a
marker in this manner. however. is not known.

 

 

 

 

 

 

 

A B
[a tan
7
..H
°t ' ta ' .a
15:0 ante 18:3 cic 6.12.14
mars . e.
1 . 1
I.“ 0.01
:d ' 0.15 ' L's =F ' 0.51 ' sir
._ 21:0 iso o'm 20:4 cis (1)6
100 a OTIS
1 ' m5 ' .a =8 eh ' .a

85

 

 

 

 

 

 

 

u»
C D
. u 14:0 iso ‘ “a, 13.0 ante
. L///”T_'_‘ . 1 /‘ "
I N 0.051
a ' eh ' ta ﬂ ' ta ' .a
liOmm un3on
0.09 0.110 .
/ /’—.———
lax‘ Cm
=5 ' ans ' 0.50 :3 ' 0.50 ‘ sis
16:1A , 17:(lcyc
0.191.‘ . o.” .
‘l . 1 A
0.0” 0.00
a ' ta ' eh i ' ta ' aa‘

Fig. 4. Selected variograms for soil community FAMEs analyzed with a minimum step of 0.07 tn (A) and 0.03 to 0.05 tn (B). and plated
community FAMEs analyzed with a minimum step of 0.07 tn (C) and 0.03 to 0.05 tn (D). The x axes are distance (to) and the y axes are

semivariancc (7).

We found a high nuggetzsill ratio for the remain-
ing 48 FAMEs, indicating that most of the spatially
structured variance in these FAMEs occurred at scales
smaller than 0.07m (Webster. 1985). In fact for many
FAMEs nugget variance approached zero [C/(Co + C)
= 1) when examined using a minimum lag interval of
0.03 to 0.05m (Table 5, Fig. 48), although the small
number of sample pairs for lag interval classes at this
scale makes such patterns statistically tenuous. It is
clear, however. that most spatially structured variance
in these FAMEs - and presumably in the microbial
communities identiﬁed by FAMEs - occurred at spa-
tial scales analogous to individual soil peds and perhaps
rhizospheres.

Whole soil vs. plated communities

On average, 20 different FAMEs were detected in each
plated community. Forty-one FAMEs were identiﬁed
in total from plated communities and only 28 of these
were also found in the soil FAME proﬁles (Table 2).
The PCA plot of the plated organisms (Fig. SA) was
signiﬁcantly different than that for the soil community
(Fig. 2A). showing that FAME proﬁles captured large
portions of soil microbial communities that were not
culturable on R2A. The two plots. nonetheless. showed
some surprisingly similar characteristics. although the

low proportion of variance explained in the ﬁrst three
PC dimensions makes this conclusion somewhat ten-
uous. We therefore reduced the dimensionality of the
data matrix for plated communities by the same means
used for soil community FAME proﬁles. When the data
were reduced by keeping only the most distinguishing
variables (Fig. 5B), the clustering of soil samples in the
ﬁrst two PC dimensions was the same as for when the
full data matrix was used, and 70 and 88% of the total
variance was explained in two and three dimensions.
respectively. Six of the seven outliers in both Figures
5A and B and the 25 soil samples forming the cluster
of stars in Figures 2A and B were all characterized as
being high in 12:0 iso, 15:0, and 17:1 ante A (Table
4). Four of the seven outliers in Figures 2A and B rep-
resent plated communities from the same soil samples
that plotted in the cluster identiﬁed by stars in Figures
2A and B. These similarities are especially surprising
given that the plated and abbreviated soil community
data sets shared only 28 of 59 possible FAMEs (Table
2). That there were only seven outlier points in the PCA
plot of plated communities (Figures 5A and B) but 25
sample points in the small cluster of stars in Figures
2A and B. and that the R2A used as a growth medium
selects against fungi. supports our contention that the
clustering in Figures 2A and B was due. at least in part,
to a difference in the ratio of bacterialzfungal biomass

86

 

 

 

110
Table 5. Parameters for variogram of FAMEs that exhibited spatial autocorrelation.
including variograms presented in Figure 4. C/(C.+C) is the proportion of population
varianceduetosparialstrucnireandA. istherange
Minimum Number of sample pairs
FAME lag (m) CJ(C.+C) A. (tn) for ﬁrst four points
Soils
10:0 0.07 0.52 0.11 29. 32. 23. 19
1 1 :0 iso ' 0.83 0.97 "
12:0 iso " 0.68 0.10 "
15:0 ante " 0.25 0.11 "
16:0 2 OH " 0.46 0.12 "
sif 9 " 1.00 0.11 "
21 :0 iso ' 0.86 0.35 "
21:1 cis 18 " 1.00 0.09 "
13:0 ante 0.03 1.00 0.08 10. 14. 13. 12
15:0 " " 0.05 "
16:0 iso " " 0.05 "
16:1 cis 9 " 0.92 0.04 "
16:0 3 OH " 1.00 0.04 "
17:0 ante " " 0.10 "
17:1 cis 10 " " 0.06 "
18:0 " " 0.05 "
l8:3 cis 6. 12. 14 “ " 0.09 "
sif 3 ' " 0.06 "
sif 10 0.05 0.95 0.12 21. 19. 27. 12
20:4 cis «.26 0.04 0.95 0.06 18. 17. 14. 23
22:1 cis 17 0.03 1.00 0.05 10. 14. 13. 12
25:0 2 OH " " 0.09 "
Plates
13:0 iso 0.07 0.45 0.17 26. 31. 23. 18
14:0 iso " 1.00 0.19 "
15:0 ante " 0.50 0.11 "
15:0 iso " 0.50 0.28 "
16:1 A " 0.77 0.20 "
17:1 iso E " 0.63 0.20 "
13:0 ante 0.03 1.00 0.08 9. 12. 12. 12
12:0 3 OH " 0.76 0.05 "
16:1 iso E 0.05 0.70 0.08 19. 17. 27. 12
sif 3 0.03 0.67 0.04 9. 12. 12. 12
sif 4 0.037 0.94 0.09 15. 12. 18. 16
17:0 cyc 0.03 1.00 0.04 9. 12. 12. 12
sif 7 0.04 1.00 0.07 16. 16. 13. 23
among soil samples. In addition. the seven outliers plated communities based on the presence of more of

(Figures. 5A and B) seemed to have a greater Gram the Gram negative marker 10:0 3 OH in the majority
positive:Gram negative ratio than did the majority of of plated communities (Table 4).

 

 

 

 

PC2

111

 

 

10:0 30H
12:0 iso
15:0 iso
15:0

 

bbhhowacaa

 

 

16:0 iso
16:0

17:1 ante A
17:0 cyclo

 

 

5

~2024681012l4

18:0

20:0

19:0 iso

Plated Communities
Plated Conununities
. Plated Communities

KH—IO'IIHIUOG>

 

 

 

Fig. 5. PCA plotandbiplots forplawd eommunityFAME proﬁles. A.PCAplot forplatedcommunityFAMEproﬁlescontainingﬂ FAMEs.
B.BiplotforplaredcommunityFAMEpr-oﬁlescontaining llFAMEsselectedusingPCA.ParamemrsarelistedinTable3.

The effect of four other FAMEs on the distribution
of sample points in the large cluster in Figure 5B is
also evident. PC 2 is a contrast of 15:0 iso and 16:0 iso
vs. 16:0 and 17:0 cyclo. That is. samples forming the
linear cluster represent a gradient of these four FAMEs.
with plated communities lying at the high PC 2 end of
the cluster being high in 15:0 iso and 16:0 iso and
low in 16:0 and 17:0 cyclo. Points at the other end of
the cluster have opposite characteristics with respect
to these four FAMEs. Since none of these FAs are
currently known to be markers. further characterization
of this gradient with respect to particular microbial taxa
is not possible.

Grouping FAMEs from plated conununities did
not reveal any further information: the same samples
which were outliers in Figures 5A and B were outliers
in the third PC dimension. All cluster analyses. again
regardless of distance metric or linkage method used.
were consistent with clustering found in PCA.

Variograms for nine FAMEs from the plated com-
munities exhibited spatial autocorrelation at a mini-
mum step lag of 0.07m. Only one of these. 15:0 ante.
also exhibited spatial autocorrelation when extracted
from soil (Table 5, Fig. 4C). That spatial structure is
revealed for some FAMEs from plated communities.
but for different FAMEs than from soil communities. is
evidence that. though laboratory media select for par-
ticular components of the soil microbial community.
the resulting community is not solely a result of selec-

tion by the growth medium. but reﬂects the distribution
- in soil - of those soil microbial populations that can
be cultured. Further evidence that microbial distribu-
tions present in soil are reﬂected in plated communi-
ties is seen in that the variograms for 15:0 ante from
botlt soil and plated communities had identical ranges
(0.11m; Table 5). That one half of the population vari-
ance of 15:0 ante was explained by spatial structure
among plated communities. but only one quarter was
explained for the soil communities, is consistent with
the fact that whole soil communities are more variable
than plated communities. Again. some FAMEs that
showed no spatial structure at a minimum step lag of
0.07m exhibited spatial autocorrelation when analyzed
using smaller minimum lags (Table 5. Fig. 4D).

Conclusions

Whole cell FAME proﬁles of soil samples offer a rapid.
inexpensive and reproducible means for characterizing
numerically dominant portions of soil microbial com-
munities. including those organisms not culturable.
By taking advantage of current knowledge regarding
marker fatty acids and using multivariate statistical
procedures we have shown that the numerically domi-
nant microbial communities of most soil samples taken
from a corn ﬁeld were similar. About 18% of samples.
however, were distinct in having lower bacterialzfungal

112

biomass ratios and about 8% of samples seemed to have
higher bacterialzfungal biomass ratios than the major-
ity of samples. By using only those variables selected
using PCA. we more than doubled the percent of vari-
ance explained by these differences. The paucity of
current information regarding individual and grouped
marker FAMEs reduced our ability to interpret similar
patterns when PCA was conducted on a data matrix
made smaller by grouping FAMEs.

Semivariance analysis may be a useful means
of identifying new marker FAMEs within particular
ecosystems. We found. for example. that 21:0 iso had
a unique range of autocorrelation. indicating that it
is a marker FAME. albeit for a currently unidentiﬁed
microbial taxa. Concurrent use of geostatistical and
FAME analyses may be a powerful means of revealing
other potential marker FAMEs.

We were able to capture and describe small-scale
patterns of distribution of microbial populations using
semivariancc analysis. Where autocorrelation occurred
it was generally at scales < 0.2 m. a scale analogous to
individual soil peds and rhizospheres. Sampling more
intensely at very small scales should result in the dis-
covery of spatial structure for more FAMEs.

FAME proﬁles of most plated communities were
signiﬁcantly different from those of whole soil com-
munities. indicating that FAME proﬁles of soils reveal
portions of the microbial community that do not grow
on R2A. A small number of distinct plated communi-
ties. however. retained the same FAME proﬁle char-
acteristics (high in 12:0 iso. 15:0 and 17:1 ante A)
that distinguished the whole soil samples from which
plated communities were cultured. As indicated by the
variograms for a number of individual FAMEs. the
distribution of a portion of the soil community that is
culturable seems to be retained when soil communities
are cultured. Thus. microbial communities grown on
laboratory media reﬂect not only speciﬁc laboratory
growth conditions. but also the distribution of cultur-
able microbial populations in soil.

Acknowledgements

We thank Lori Merrill, Deane Lehmann. and Chenn-
Ching Chou for sample collection and laboratory assis-
tance. and Helen Garchow and Peter Stahl for advice
regarding FAME analyses. This project was supported
by the National Science Foundation (NSF) via grants
to Michigan State University for the Science and Tech-
nology Center for Microbial Ecology (BIR9120006).

88

the Long Term Ecological Research Project in Agri-
cultural Ecology (DEB9211771). and a Doctoral Dis-
sertation Improvement Award (DEB931 1380); support
was also provided by the Michigan Agricultural Exper-
iment Station.

References

Applied Biostatistics. Inc. 1992 NTSYS. Exeter Publishers. Setauket
New York. USA.

Bobbie R J and White D C 1980 Characterization of benthic micro-
bial community structure by high-resolution gas chromatography
of fatty acid methyl esters. Appl. Environ. Microbiol. 39. 1212-
1222.

BnrssaardLBouwmanL.GeursM.HassinkJandZwartKB 1990
Biomass. composition and temporal dynamics of soil organism
of a silt loam soil under conventional and integrmd management.
Neth. J. Agric. Sci. 38. 283-302.

Digby P G N and Kempton R A 1987 Multivariate Analysis of
Ecological Communities. Chapman and Hall. New York. USA.

Erwin J A 1973 Fatty acids in eukaryotic microorganisms. In.
Lipids and Biometnbranes of Eukaryotic Microorganisms. Ed.
.1 A Erwin. pp 41-143. Academic Press. New York. USA.

Everitt B 1980 Cluster Analysis. Halsted Hess. New York. USA.

Findlay R H and White D C 1983The effects of feeding by the
and dollar Mellita quinquiesperforata (Leake) on the benthic
microbial community. 1. Exp. Mar. Biol. Ecol. 72. 25—41.

GillanFTand Hogg R W 1984 A tnethod fortl'ie estimationofbac-
serial biomass and community structure in mangrove-associated
sediments. J. Microbiol. Methods 2. 275-293.

Harwoodl LandRussellNJ 1984 LipidsinPlantsandMicrobes.
Allen & Unwin. London.

Jantzen E and Bryn K 1985 Whole-cell and lipopolysaccharide fatty
acids and sugars of Gram-negative bacteria. In Chemical Methods
in Bacterial Systematics. Eds. M Goodfellow and D E Minnikin.
pp 145-171. Academic Press. London.

Jolliffe IT 1986 Principal Component Analysis. Spdnger-Vedag.

Klug M .1 and Tiedje l M 1993 Response of microbial communities to
changing environmental conditions: chemical and physiological
approaches. In Trends in Microbial Ecology. Eds. R Guerrero and
C Pedros-Alio. pp 371-374. Spanish Society for Microbiology.
Barcelona. Spain. .

Liu C J and Keister T D 1978 Southern pine stern form deﬁned
through principal component analysis. Can. .1. For. Res. 8. 188-
197.

MayberryWR.LambeDWandFergusonKP l9821dentiﬁcation
of Bactemt’des species by cellular fatty acid proﬁles. Int. J. Syst.
Bacteriol. 32. 21-27.

Microbial 1D lnc. 1992 Microbial Identiﬁcation System Operating
Manual. Version 4. Newark. DE. USA.

Milligan G W and Cooper M C 1987 Methodological reviews: clus-
tering methods. Appl. Psychol. Measurement 11. 329-354.

Moss C W 1981 Gas-liquid cinematography as an analytical tool in
microbiology. .1. Chrom. 203. 337—347.

Moss C W. Bees 5 B and Content G O 1980 Gas-liquid chromatog-
raphy of bacterial fatty acids with a fusedosilica capillary column.
J. Clin. Microbiol. 12. 127-130.

Perry 0 J. Volkman J K. Johns R B and Bavor H .1 1979 Fatty acids
of bacterial origin in contemporary marine sediments. Geochim.
Cosmochim. Acta 43. 1715—1725.

Robertson G P and Gross K 1994 Assessing the heterogeneity of
belowground resources. quantifying scale. In Exploitation of
Environmental Heterogeneity by Plants. Eds. M M Caldwell and
R W Pearcy. pp 237-253. Academic Press. New York. USA.

SAS Institute 1991 SAS User's Guide Version 6.03. Ed.SAS Inst.
Cary. NC. USA.

Torsvik V. Goltsoyr .1 and Daae F L 1990a High diversity in DNA of
soil bacteria Appl. Environ. Microbiol. 56. 782-787.

Torsvik V. Salte K. Sorheim R and Goltsoyr J 1990b Comparison of
phenotypic diversity and DNA heterogeneity in a population of
soil bacteria. Appl. Environ. Microbiol. 56. 776-781.

Trangmar B B. Yost R S and Uehara G 1985 Application of geo-
statistics to spatial studies of soil properties. Adv. Agron. 38.
45-94.

Vestal J R and White D C 1989 Lipid analysis in microbial ecology.
Bioscience 39. 535—541.

Vrljoen BC. .1 L F Koch and P M Lategan 1986 Fatty acid com-
position as a guide to the classiﬁcation of selected genera of
yeasts belonging to the endomycetales. J. Gen. Microbiol. 132.
2397-2400.

89

113

VolkmanJK.JohnsRB.GillanFT.PerryGlandBavorHJ
1980 Microbial lipids of an intertidal sediment - 1. Fatty acids
and hydrocarbons. Geochim. Cosmochim. Acta 44. 1133—1143.

Webster R 1985 Quantitative spatial analysis of soil in the ﬁeld. Adv.
Soil Sci. 3. 1-70.

White D C 1983 Analysis of microorganisms in terms of quantity
and activity in natural environments. In Microbes in Their Nat-
uralEnvironmemsEds.JHSlater.RWhittenburyandJWT
Wimpenny. pp 37-66. Cambridge University Press. Cambridge.
UK

Zelles L and Bai Q Y 1993 Fractionation of fatty acids derived from
soil lipids by solid phase extraction and their quantitative analysis
by GC-MS. Soil Biol. Biochem. 25. 495—507.

ZellesLBaiQYandBeeseF 1992 Signaturefattyacidsinphospho—
lipids and lipopolysaccharides as indicators of microbial biomass
and community structure in agricultural soils. Soil Biol. Biochem.
24. 317-323.

CHAPTER 5
SUMMARY, SYNTHESIS, AND FUTURE RESEARCH POTENTIAL

Microbial ecology is a broad ﬁeld that encompasses many disciplines, from the
subcellular (molecular biology) and the cellular (microbial physiology), to the ecosystem
(biogeochemistry), the landscape (landscape ecology), and the entire earth (earth systems
ecology). Microbiologists have relied heavily on studies of the physiology of
microorganisms isolated from their environment to understand the regulation of microbial
processes (e. g. Conrad 1996). Many of these processes are fundamental to ecosystem-
level functions, such as decomposition, nitriﬁcation, denitn'ﬁcation etc. Biogeochemists
and other process-oriented scientists, on the other hand, study these same processes but
often do not explicitly consider the microorganisms responsible for those transformations
(6. g. Conrad 1996). Instead, they focus on measuring substrates, products, and the
environmental conditions under which transformations occur. Both methods have their
advantages and limitations. Together, they can provide complementary information that
can help lead to new understandings of nutrient cycling and microbial ecology. In this
dissertation I have borrowed from the techniques of both microbiology and
biogeochemistry to address two fundamental questions in microbial ecology: 1) is
microbial diversity ﬁrnctionally signiﬁcant, i.e. what are the ecosystem-level
consequences of microbial diversity, and 2) how are microbial communities spatially
distributed.

In order to focus the question of whether microbial diversity is functionally
signiﬁcant, I chose to study nitrous oxide (N 20) production by soil denitriﬁers. I

sampled soil from a conventionally-tilled agricultural ﬁeld and a never-tilled successional

90

91

ﬁeld. Results, using a soil enzyme assay (Chapter 2), show that oxygen inhibited the
activity of enzymes involved in N20 production (nar, nir, nor) to a greater extent in the
denitrifying community from the agricultural ﬁeld than in the community from the
successional ﬁeld. The nar, nir and nor enzymes of the denitrifying community from the
successional ﬁeld, on the other hand, were more sensitive to pH than were those in the
denitrifying community from the agricultural ﬁeld. Moreover, the denitrifying
community in the soil from the successional ﬁeld had relatively more active nos
enzymes, which reduce N20 to N2, than the denitrifying community in the agricultural
ﬁeld. Also, the rate of change in the relative rate of N20 production with increasing
oxygen was different for each denitrifying community. Each of these differences
suggests that the denitrifying communities in these two soils are different and that they do
not respond to environmental regulators in the same manner.

I also isolated 156 denitrifying bacteria from the same soil samples used for
enzyme assays in Chapter 2. Using fatty acid methyl ester (FAME) proﬁle and cluster
analyses, I classiﬁed these isolates into 27 different taxa. Community structure, based on
this classiﬁcation, was different in the two soils. I then characterized the activity of their
nos enzymes, which are responsible for the reduction of N20 to N2 during denitriﬁcation.
There was substantial diversity in the degree of nos sensitivity to increasing oxygen
among denitrifying isolates. This physiological diversity is biogeochemically signiﬁcant
and demonstrates a clear potential for differences in denitriﬁer community composition to
affect differences in N20 production among ecosystems, independent of direct
environmental controls. Chapters 2 and 3 of this dissertation provide some of the ﬁrst
evidence that microbial community composition can inﬂuence ecosystem function, and

speciﬁcally that soil denitriﬁer community composition can inﬂuence soil N20 ﬂux. At

92

least for N20 ﬂuxes, it may be necessary to consider microbial community composition
as an additional factor inﬂuencing ﬂux rates.

Although spatial variability of soil microbial community structure has been
addressed at relatively small scales using isolates and microscopy, little research in this
area has been conducted at the ﬁeld scale. In addition, soil microbial community
structure has rarely been quantiﬁed. Chapter 4 (published in Plant and Soil, 1995) was
the ﬁrst study, to my knowledge, that quantiﬁed the spatial variation of soil microbial
community structure at a ﬁeld scale. Applying both principal components and
geostatistical analyses to FAME data extracted from soil samples, I was able to show that
soil microbial communities are largely similar across a conventionally-tilled corn ﬁeld '
and that differences in community structure, where they existed, were most likely due to
different fungal:bacterial ratios. In addition, the distribution (range of autocorrelation) of
many fatty acids mirrored the distance between corn plants within the row along which
soil samples were taken. It has since become common to use FAME analysis of soil
samples and multivariate analyses of these data to characterize soil microbial community
structure (e.g. Frostegaard and Bath 1996, Sundh et al. 1997, Saertre 1998). I am not
aware, though, of any further work using geostatistical analysis of FAME proﬁles,
despite the power of this technique to quantify ecologically relevant spatial variation (6. g.
Robertson and Gross 1994).

A number of future research avenues are indicated by the research presented in
this dissertation. For example, as suggested in Chapter 2, it could be that denitriﬁer
community composition may affect not only denitriﬁer enzyme activity, but also
induction. Denitriﬁcation in situ is highly temporally variable (e. g. Christensen and
Bonde 1985, Sexstone et al. 1985a, Christensen et al. 1990b, Christensen and Tiedje

1990). Thus, the status of the denitriﬁcation enzyme induction at the time of

93

denitriﬁcation events (largely in response to rainall or irrigation events) can play an
important role in determining the NZO/(N20+N2) ratio (Dendooven and Anderson 1994).
It could be that different denitriﬁer communities respond to the same environmental
regulators at different rates so that, for example, different communities produce N20 for
longer or shorter periods of time following a rainfall event or that different communities
maintain their denitriﬁcation enzymes following a dry period for different amounts of
time. The enzyme assays described in Chapter 2 could be applied readily to address such
questions.

Further study of the physiological diversity among denitrifying isolates collected
from the two native systems could also provide further information on the importance of
denitriﬁer diversity to N20 (and NO) production. The control of induction and activity of
all four denitriﬁcation enzymes requires further study as indicated by Chapter 3 and other
recent publications (e. g. Komer 1993, Robertson et al. 1995, Conrad 1996, Ka et al.
1997). The 156 denitrifying isolates I collected for this chapter have been stored in
glycerol and are available to interested researchers.

Finally, spatial variability of soil microbial community structure is an important
link between the three research reports in this dissertation. Denitriﬁcation and probably
denitriﬁers, have notoriously high spatial variability (e.g. Parkin 1987, Christensen et al.
1990a). This variation needs to be quantiﬁed in order to better predict soil N20 ﬂuxes.
FAME analysis of whole soil extracts is probably not a viable means of addressing these
questions since denitriﬁers represent < 5 percent of the total soil microbial community
structure (Tiedje 1988), but geostatistics could be applied to soil samples collected in at a
relevant spatial pattern and scale and analyzed using the type of soil enzyme analyses
described in Chapter 2. Spatial and temporal patterns of both denitriﬁcation enzyme

induction and activity could be studied using these tools. Relevant scales of investigation

94

could include within and among aggregates (Sexstone et al. 1985b, Seech and
Beauchamp 1988), at different distances from decomposing residues (e.g. Parkin 1987,
Ambus 1996), across oxidation/reduction gradients (e. g. Rice et al. 1988), and across
ﬁelds and landscapes (e.g. Grofﬁnan and Tiedje 1989, Robertson et al. 1997). Such
research could help further our understanding of soil N20 ﬂux, speciﬁcally, and soil

microbial community structure and function, in general.

APPENDIX

95

Table A1. FAME proﬁles for 35 reference strains and 156 denitrifying bacteria isolated
from soils from the conventionally-tilled agricultural ﬁeld and the never-tilled
successional ﬁeld at the KBS LTER site.

Note: Taxon and isolate numbers are as in Table 3.3. FAME proﬁles were generated
using the aerobic procedure in MIDI (1992). Values are percent of total fatty acids for
each isolate. FAME nomenclature is that used by MIDI: the number of fatty acid carbon
atoms is indicated by the number before the colon and the number aﬁer the colon
indicates the number of double bonds. A number following an “co” refers to the location
oa a double bond relative to the ester end of the molecule. Cis and trans conformations
are noted with a “c” or “t” following this number. No letter is present if the conformation
is unknown. “Ante,” meaning anteiso, and “iso” mean that a methyl group occurs at the
third or second carbon from the ester end, respectively. “10 Me” means that the 10th
carbon from the ester end is methylated. “Cyc” refers to the cyclo propane analogue. The
number before OH refers to the location of a hydroxy substitution relative to the carboxyl
end of the molecule. A capital letter at the end of a monounsaturated acid indicates that
the position of the double bond is not known but is at a different location than for other
monounsaturated fatty acids of the same chain length and branching pattern but different
letter designation. “Sf ’ means “summed feature” and indicates that more than one FAME
has a particular retention time on the gas chromatograph (GC) column (in minutes).
Speciﬁcally, sf 3 is some combination of 12:0 ald (fatty acid carboxyl end substituted by
an aldehyde), uk 10.928, 16:1 iso 1, and/or 14:0 30H; sf4 is 16:1 m7c and/or 15:0 iso
20H; sf5 is 17:1 iso and/or 17:1 ante B; sf6 is 18:2 (06, 9c and/or18:O ante; sf7 is 18:1
w7c, 18:1 (09c and/or 18:1 0) 12t. A number following uk refers to the retention time of an
unknown compound.

Table A1.

 

96

IL
31111lIllI1|lllllllllllllllllllllllll
a
F
fllllll11lllIllllllllllllllllllllllll
F
8

3

140

 

Fattyaeld(percent)
12:020H 12:130H uk13.566 12:0301-1 1420190

 

10:03OH uk12.112 12:0 13:01.0 13:0anfe

sassts
g $39999
,~maa888889983383ﬁ£ﬁ88353§§§3§gzzzzeﬁ
i resist
’ §§¢¢ur
gv-v-v-v- FFFFFFF '- FFFFFFF v-PI-v-v-v-v- FFFFF PPFNN
p.

 

97

 

I I l 1 1 1 I md
I II I II N.” Is I |

0:8 to. 0393 0.0206. rel—”ow 0.0058 ﬂow 99 003 Km. 859m—

 

?889 38 E:

0% 1 1
o.— 1 l

8.99 8m.v—x: (OE-pum—

001 pnmw

 

8.
9
629 : as .a
629 : 25.1
5.9 325.1
:99 : 85 .a
:3. .325 .23
Sea noes .8<
m:
E
3.
8.
3
a
3
8
8
2.
E.

N
[s

50198888899988888
aFPFPFFv-PPFFPFFPFFFFPPFFFFFPFFFPPI—PNN

S

8953:8300.

Assamese .2 22:.

98

 

8.2 .5 (8:. t:

022. one

880?.- ﬁt

 

3 ..
3 r
u no
838. :2 8.212..
€885 :8 as“.

:00 on. our;

90 '

one how one to— 88 to—

 

on.
a.
820. : RE .1
629 : as .a
619 : as .a
:29 : 82.1
:5. cause .23
.33 e8: :3.
m2
5
3.
8,

501288888939838881‘33288353

.363: 833. :95.

 

Q8583 .2 case

99

 

ll m.° I

one to. one :2 00.90..

109 one—

Oontow 1906— 180190.

 

E§§¥§€u

out.

md

£35

1 r t ed

008 t: 008 t: cacao“: 00.0”:

 

8.
2
§EIIS¢
.86:;S¢
§EIZS¢
c99:8?a
.ieﬁiﬁég
.Qﬁeﬁsﬁ<
R.

.3
3.

8.

3
a
3
8
8
2
R

N
h

~m98888893988888

89.5: 80.3. :96...

€8.28. .2 as...

100

 

FM

8.08 «new

a... 1
Nu v.—
VN 1
ad 1
o... I

catnip. ﬂow 02 o— 06-

1820—. ooagoonmp 8:092. 00.99

«.mp
ad—
VN
Nd
...m
Wm
of
NO
md
a. v
5.—
ON
ON
md
m.—
#8
v...“
mN
EN
mi

 

€88. :8 at“.

019...— IOMOK— IONOH: 020.90.

 

on.

a.
620. .. 25.1
.86. .. as .a
5.0. .. as .a
5.0. .. 2:. .a
5.0. .325 .8?
.8... .38.: .3
r...
...
3.
8.

#01918888899983888531‘2288353

895.. 80.8. :98»

.8828. ._< 2%...

101

 

 

8.3 ... n .. o...
.8 u .. .. ..o
02 .. .. .. a.»
.8 .... .. .. n...
....3 .. .. .. a...
«.2 .. .. .. no
n.2, .. .. .. o.«.
«.2. r .. .. as
v.8 .. no .. E.
0.8 .. n .. 3
m8 .. .. .. on
«.R .. .. .. as
v.5 .. .. .. v...
8.8 .. .. .. no
:8 .. .... .. a...
28 n I .. «.o
..k .. .. .. o...
no. .. .. .. ..m
3:. .. .. .. mg
a? .. .. .. on
...... .. .. .. no
o... .. .. .. o...
02 .. .. .. ...a
08 .. .. .. «.m
...... .. .. .. a...
«.8 .. .. .. to
3.. .. .. r n...
«.«« .. r .. m...
8.2 no .. .. ..m
an: .. .. .. .3
«.8 .. .. .. «.o
...... .. .. .. «.o
8.2 .. .. .. «.v
mm. 1 .. .. n...
n.«« .. .. .. ....
«:0 8.0 8.0 I.» 8.0
88.8.3823

 

8.
«.
8.0. .. as .a
8.0. .. 8.. .a
6:0. .. a... .a
5.0. .. as .a
....0. .325 .83
.«a.« 88.... .3
2..
...
3.
8.

~a988888989888888‘lﬂf388333

g3 ugh

 

A8288. .2 as:

102

 

“—00.29 await... 9'—

...—

vé

v.0

No
#6

00.9: 10092 89033 100 ﬁNw ION9N- 8:099 00.92 9N. «SN—.3 10090.

 

.288. 28 ...-u

 

hop

5
NNNNNNNMDVQDIDIDIDIDIDIDIDIDIDIDIDOOOODDODDDDDOD

.8288. ._< 2%..

103

 

1 s. u —
1 0.0.
1 90

1 0.0—
1 0.0.

1 0.0.
1 0.0
I 5.0—
1 0.0—
1 0.N..
1 0.0..
1 0.0..
1 No—
1 0.0
1 5.0

1 mm—
1 9s.
0.0 Na
1 «.9
1 0.0
1 v.2.

038—0.. 00.90. 0.0290— 18.—”0.. 0.005820. 90' 008 t0. 8:090— 00.90—

0.0
0N

0.0
9 ..

 

9.8.8. 28 Eu...

nNN
h. pN
N... w
New

9N0

90
0N
w.—
MN
9'
0.0
h...
0..
9 w

000.: x: (8:0 $0.

000. Km.

 

b0.

50—

y.

0

P
NNNNNNNM1')V1010IDIDIDIDIDIDIDIDIDIDOQDOODOOODOODD

...N p
808:: 80.0.. :98»

 

.8288. .2 28..

104

 

00.0— .3 < 8:0 put

02 0— 90— 008 8:0 ...:

00800. ﬁt

 

.288. 28 ...-a

00800. t:

100 00. 90..

0.0 w
0.0 p
0.0.
9!.
p.0 _.
0.0 v
90 w
90 p
EN ..
0.3.

0.0—

0.0
0.0.
0.0—

0.0..

NS
90*
ed—

5.0
90'

008 K0. 0h8 t2 008 t0.

 

3.—

E
NNNNNNNNMV’IDIDIDIDIDIDIDIDIDIDIDIDDDDODDODODODDD

R—
353!!! :98»

.88....8. ._< 28...

105

 

008 $0.. 008 p.05 00.90.

100 90'

001—”05 ...ON90.. £00390?

 

€88. :8 and

95..

I I I 0.0 0 ..
I I I 0.5 N N
I I I 0.5 0 w
I I I 0.0 0 w
I I I 0.0 0 N
I I I 0.0. 5 0
I I 1 N0 0 N
I I I 0.05 0 0
I I I 0.05 0 0
I I I v.0 5 ..
I I I 0.0 0 0
I I I 0.5 0 v
I I I N.5 0 0
I I I «.0 0 0
I I I 0.0 0 w
I I I 0.0 N 0
I I I 0.0 0 w
I I I 0.0 N p
I I I 0.0 0 N
I I I 0.5 0 N
I I I 5.05 0 0
I I I 0.: 0 N
I I I ... 3 0 N
I I I 0.0 v N
I I I 5.0 0 N
I I I 0.0 v N
I I I vN 0 w
| | 0.0 F. P I

 

..3

E
NNNNNNN00000000000000000000000000000

i1

.8288. .2 28..

106

 

.... I I
o... I I

II “.0 II
00.08 NUON 09.0.08 0.0N 82 0— 90¢

:83. 83892 283. 830.

0.0. I
0N5 I
9N— 0.0
.0 I
0.0 I
N0. 1
—.v_. I

 

€88... 1823.

00.95. ...0095— ION95.. 050590.

 

50..

S;
505
0N..
VNF
0:.

E
NNNNNNNDMVIDIDIDIDIDIDIDIDIDIDIDIDQODDDDDDOCOQDD

5N ..
.305: 20.00. 098..

08:88. ..< 28..

107

 

0. .
..N
0..
0.0
0..
0..
1..
0. .
5..
...
0..
MN
0.:
0.00
0.00
5N0
0.0
0.00
0.05
0.00
0.00
..00

50.0 0%.0 00.0 :00 8.0

 

€8.80 28 E...

 

8.
50.
50.
.v.
50.
mm.
VN.
0..
5..
0..
0
3
.o.
00
Na.
00.
00.
55.
3..
0m.
5m.
35:: 0.0.00. .5qu

NNNNNNNM('0VInIDIDIDIDIDIDIDIDWWWDDODODDQDDOOOO

 

ABEESV .2 use.

108

 

0..

... 3. .0. 00nd. .3

Nu
v.5
0.0
5. .
ON
ON
ed
90

5.0

0.0
V. .
N. .

Y.
...
9.

9?.

39?. ...-0093 80.0.... IOn.N. IONON. 01.90.

 

‘

€88... 38......

00.90. 9N. N...N..3 10090.

 

8 v.
.5 v.
a w.
an ...
«a ...
on. a.
3. n.
8. u.
8. a.
no u.
8. ..
3. ..
«a ..
8. a.
8 a.
no. a
8 o
.86.... ...-m .
5.9.... ...-m s
a: s
8 ..
8 s
t. .
a... s
a. s
3 .
an s
.n s
an s
R .
8n 0
«on o
.8 o
8. o
2.. o
8. o

g; E...

€2.28. .2 2......

 

 

109

I 0.0. 9. 5.. I 0. . I 0N. Nun I 0.. I 00 v.

 

I a... .... .... .... a . I ...... ....u I .... I ... v.
I .... ... ... n... N. I v... ...... I , v... I cu v.
I v... I ... I .... I m... Nun I o... I 8 c.
N. a... o... ... ... N. I .... New No I I «a v.
I o3 .... ..m m... I I ..m. o... I I I 8. n.
I o... ...... ... m... I I a... o... o. I I 3. n.
to ....N m... «N I I I o... ... . I a a I 8. u.
I o 8 I ...... I I I m... to no I I 8. u.
I a R I Nu I I I .... o o I I I w. a.
I N 8 I ... I ... . I ta... . a I o . I 8. ..
I v .. m N a . I I I ...... a .. I .. . I .... ..
.. .R I .. on a. I ..«N .u. I I I 3 ..
I o m a o I I o . I ....8 v nu I I I 8. c.
I n ... I I I I I New N... I n . I 8 o.
I o . I I I I I b ca n... I m o I a... a
I n... I .... I ... I ..8 v... no I I 3 o
I ...... m... R I I I N... N... I v...” I .86.... ...m s
I ..o. m... N. I m... I v.8 .... I ... I .86.... ...m ..
I .... N... N. I I I 3. ..o. o... I I a: h
I .... ... ... I .... I .... N .. .... I I 8 .
I ... .... ..N I I I a... v... ..o ..m I 8 .
I .... m... Nu I a... I v.8 v... I ... I t. h
I ..~. ..o .... I m... I v.8 no .... I I on s
I ... .. I ..~ I I I ...... .... ... I I n. .
I N .. N. ..N m... I I N8 o... I I I 3 ..
I I. Z ...N .... I I N8 ..o. .... I I N. .
.... ... .... ..N ...o I I N .. o. .. ..o 3 I .n .
I ... o. 2 .... I I ..m. .... m. ... I an ..
I m... ... ... . I No I ...... a... I ... I R .
I .. .. I o... I ..o I Nmu ... .... I I 8. o
I v... N. 3 I I I ....Nu o. m. I I «8 o
I m... .... ...... I I I ..8 N. ... I I .8 o
I m... ..~ ..~ I I I .8 ... I I I 8. a
I .... I .... I I I ...n .... ..o .... I a... o
I. .... .... .... ...... I I ...... .... n. m... I ...... .
...... .6. 8.9.. 2.29... :8. .... 2...... .... 9.... 8a ...... 259.... 8.9.... 8......3 <2... .... 68. ...... 82.5.58. :9...
6.2.... :8 ......

 

 

9.25:8. ..< .3...

110

 

00.0. .3 (0.:- .“5.

0.0

..N
020. 90.

8a ...... ....

00800. .H5.

 

.28.... so. a...

0..
00800. ..5.

100 00. 90.

0.0
9N.
..v.
0.0.
0. . .
0*.
5.0.
90.

008 .n0. 058 .0. 008 .H0.

 

8 ...
0. ..
8 ...
8 v.
8 v.
8. n.
3. ....
8. a.
8. u.
8 a.
8. ..
.... ..
um ..
8. o.
8 o.
8. m
o. o
.86.... .80 ..
.86.... ...m ..
8.. ..
8 .
8 ..
t. ..
8 n
8 .
... .
8 ..
... .
8 .
8 .
.8 o
«8 o
.8 m
8. o
8. o
8. o

806:: 80.3. :98...

828:8. .2 .3...

111

 

008 ..0. 008 .0.

..N

1.0
0.0
02 90.

...-00 90.

002.0. ION90. 10000.90.

 

88.2.. no. ......

...
N..
0.0

95.

0>095. 008 .5. 008 .H5. 3:095. 00.95.

0. . .
0.0
0.5
0.0
0.0

...

 

8 ...
.... c.
8 ..
8 v.
8 ...
8. n.
8. n.
8. u.
8. N.
a. a.
8. ..
3. ..
«m ..
8. o.
8 o.
8. m
8 a

.86.... ...-m ..

.86.... ...m u
a: ..
8 .
8 .
t. ..
8 ..
8 ..
....” ..
8 ..

.n .
8 ..
R .
08 o
«8 c
.8 m
8. m
8. a
8. o

.305: 838. 098..

885...... .2 .3...

 

112

00.08 Non 0N ..0.08 0.00 02 0. 90.

' '.0 II II
II II o. P I.
ll II n. F II
II In. n.° II

I I 0.0 0.0
18.0. 00805090. 00690. 00.90.

 

ll ”.0 II II
I I I 2.
I.- h. F ll II
II '.° II II
II “.0 II II

00.95. 10095. 10095. 020.90.

5.0
0.0
0.0
0.0
0. .
0.0
0.0
0.0

0.0
5.0

0..

90 .

 

.88... 28.5....

.86.... .80
.86.... .20

 

82.8.8. .2 2......

 

113

I I I I I N. ..
I I I I I ..N ...
I I I I I 8 v.
I I I I I 8 v.
we I I I I 8. n.
.. I I I I 3. n.
N. I I I I 8. N.
... I I I I 8. N.
n. I I I I 8 N.
I I I I I 8. ..
.... o... .... I I 8. ..
I I ..N I I N... ..
I I I I I 8. o.
n. I no I I 8 o.
m. I I I I No. .
I I I I I 8 o
I I I I I 86.8.80 N
I I I I I .86.... .8 N
I I I I I N.. N
I I I I I 8 N
I. I I I I 8 N
I I I I I NN N
I No I I I 8 N
I I I I I 8 N
I I I I I 8 N
I I I I I N. N
I I I I I .n N
I I I I I 8 N
I I I I I NN N
... I I I I .8 m
m. I I I I N8 8
I I I I I .8 m
I I I I I 8. o
8.. I I I I 8. .
... I a... I I 8. 8
NE... 8.... 8...... 32m 8.. 3.5.2.8. ceaN

 

 

.88.... 28......

 

.8288. .2 ......

 

114

I I I I I I I I I I I I I N...
I I I I I I I I I I I I I 8.
I I mo I I ...N I I I I I I I v.
I I I I I I I I I I I I I ..N.
I I I I I I I I I I I I I NN.

88888888888888888888888

I I N.. o... I I I I I I I I I o. o.
I I I ..N I I I I ..N o... I I I 8 N.
..N I q. n. I I I I I I
m... ... .... ..N I I I I I I I I I o... m.
N.. I ... a... I I I I I I I I I seasons... v.
...m ... a... .

a... I ..

«N ..N ... 3 I I I I I I I I I 8 ...

.... N... N. .

N... N. o.

n.» I N.. ..N I I I I I I I I I 3.. v.

o... I no no I I I I I I I I I .m ...
“8.3. 88.1.... 3. 8.3. 1028. cone... :88. 1038. 883. 8.3. ..N. N...N...: 102.5. 352.18. 85..

 

 

€88... :8 ...-n.

 

€3.88. ...... 2......

115

 

I ...2
I 0.9
...p of.

0N
MN
n. F
m. w
ad
a. —
9N
c.—

h. _.
«N
N...
0..
v.0
—.N
5..
ad
PM
Né

md
b. w

ad
N.
0..
o.—
v.—
o.—

o.—8 «6.. 0390— 3-299 18:6. Boone to. cum—

 

né

....
....
0.—
ad

N.
N.—

83 ..m. 3:. 9m. 8. cum.
€BRF§§E

0.0
v.0.
0.:
«.9
mNN
ad.
adv
m6—
v.2
0.0.
v.0.

ndn
v.8
Q3
Won
ad
0.»
0.9
adv
n.0—
EN
5. ..N
Nan
Nap
v.00
v.3
..KN
him

I n;
I Q.
0.. I
0.. I
I md
I we.
0.0 I

8m.: x: < 8:- pump

00.. imp

 

8w
n:

how

88888888888888888888888

8. m.
.85....Sd v.
8. v.

.o v.

53882
3.

895.. 853. :92.

€8.88. .2 2.5

116

 

$6. a: (as. t:

02999

En
vd
..N
ON
..0
—.N
ad
5..
v.0
88 8:0 ts.—

omaoo. PK.

 

€8.23 28 ...-h.

0.?
ad—
ad
ad
0.....
Nd
van
0.0
EN
«.5
88 on. t:

IOn 00. 9m.

90..

I .69
I ‘8
I O60
I 0.8
N N ed
..N 06
co m.m
a w 0.0
«N ad.
5.. GS
a.— m6.
NN v.2
a p 06
a p 0.0
o w No
a _. a.»
m — 0.0
N. .. 9.0
n F —.o
m P Rd
N. _. md
0 _. m6
5 p ad
I 06
I ON
I v.—
I v.0
I n.—
I m6
I a;
I n;
I m6

88 to. one to. 88 to—

 

88888888888888888888888

E83. deli a.

o. o.
8 N.
Nu m.
ow. m.
gagged ...
8. ...
.o v.
2 ...
«N v.
8 v.
3. v.
.m v.

.882. 313. :98...

 

€25.28. .2 2%..

117

 

I 5.0 |
I I ON
I h... né
I 0.0 I

one ..uc— 88 ﬂow 00.90—

18 92.

08:6. IONono— 18010qu

 

€32§§€N

on:

ms

gee“: 008 th— 88 t: 08.09: 8.0.x...

«.9
a.»
m.m
md
9m

v.0—
0.»
EV
ed
N.m
V...»—
v...
NS

Nm

 

88888888888888888888888

o. o.
8 N.
Nu m.
o... m.
.SBXKzi ...
8. ...
.o ...
oN ...
NN ...
8 v.
..m ...
.m ...

.383: 238. .8qu

835.83 ..< 2.3

Table A1. (continued)

 

 

Fatty-OHM)

18:010Mo 17:020H17203OH 17:0!80 19:0bo 1920mm 19:0cyCm86 18:120H

 

19:0 10 M0 20:3 06.9.12: 20:2 06,9:

18:0

Isolate numbu

Taxon

0.4

0.7
1.2

51
54
63
73
76
81
168
P.1ype3(e1oa)
140
62
as
10
11
13
1e
7
a
9

H8

59
38
121
42
11.8
109
63
56

ID
111111110

333333339232:999288888888888888888818113888

 

Table A1. (continued)

 

“wwﬂwwﬂ**-

 

H9

§§111111112111111111131I111111111§§§§
gIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
U)
gunuuzzusugnu:33232222332333:3322unuu
gIII1II1111$1111I1111111111111111112:
gIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
(I)

§ E
E 6 s
=538223§E§$s2;:2222283a3882§8§3§§883§
i I §
gzzzzzzzzeet2288882882aaaaaaaaamannnn

 

120

 

I I md

man—”mp await... on:

I m6
I ﬁn
I EN
I Nm
.1 v.0
I ad
I MN
I ON
I ...N
I ad

09.0”: 10095 809.3 IOn «NP 199.1% 0:99 8.90.. 99. NSNCS 10092

 

3:022: 20- En“.

Nd
vd
m6
0m
m6
ON
md
No
0.0
a. —
56
n6
ON
md
w. —
v.—

w...
ON
m6
ﬁn
md
Np
m. w
ho
ad
ad
m6
0.0
96

 

2
ER .83 .1
«5.. 5.3.1
.9... Egon .m
85. .359...- .1
892.. Even m
83.... Evan m
892.. Even .m
8923 Esau m
82a :58 .m
89... Evan m
.Btiim
ogwiim
88 o .i .1
9a .a .m
23.. .511
m3: .5- .1
BE a i .m
«8: < .8 m
9:
a.
n:
8
3
5
an: u. .3 .m
~58 .2. .m
.2
.8: o .8 m
82.0.8.1

g8888888$3§§§RRRRNRRRRRRRRRRRRRRRRRRQS

h

Gosécoe .2 23

Table A1. (continued)

 

 

Fatty acId (percent)

15:11.06 15:1 antoA uk14.503 15:01.0 15:00:11. 15:1 (1266 15:0

 

16:0 ho 16:1m11c

16:11:17ch 16:1IaoI-I 16:0Ndc

O

166

180

161

182

183

166
P.ﬂu.C 17400
P ﬂu C 17561

gnnnnnnnn

121

5.5

v- 5" :-
§§ ﬁ§§§g§§§§§§§§§§2§a
§.:5382§§;;::ga"§§5§§§§§”'2

a ..33 .. a:
«:3 sinus-Eiﬁiﬁéﬁggii
anannannannaaaaaaaaaanaaaaas

 

Table A1. (continued)

 

 

Fatty acid (percent)

17:1 18009::

 

17:1entem9c 16:010Me 17:1-mu uk16.56

17:1 b00050

15:0 180 30H

eeeeeeeeeeeee

9 R 5 a

8

Y..lllllilllli

o

p

g

SQNQQ‘TNQQ'QQQ
v

383 8 813-101

v

8

T..llliillllll

o

w

" N9
‘3 3E 33
§§§§§§§oo§.u
g 33.2 3.5.!
_ mm “x

ganannaa xxx

92

eeeeeeeeeeeeeeeeeeeeeeeee

8 8 5 a N 232 8595
11111111111111111111131§I
0.".YVQNNQYﬂQﬁﬁﬁﬁNQQNYQQYQQ
aaaeeaaeeaueaﬁ988 5R33§99
IIIIIIIIIIIIIIIIIIIIIIIII

§§§§a§§§§§§§§§§§§§
BSSE§§2522g5~”EEEEEE§§"E

--aa.'

Simmliiiii§§§§§§§§

IQRRQRRQRRRRRRRRRRRRRRRﬁRS

 

Table A1. (continued)

 

 

PM add (P070001)

17:0 16201803011 16102011 16311806 16:030H 16:01.0 16210090 162111150

 

V
111111.41

9
0
9
7
2
7
2

2
2
9
7
1
1

2
6

17:01.0 17100010 1721080 17:1 «160 17:00y0

2 ES
§§§§§§§55

a:
- ll.
ganannnnn

17.4
6.9
13.7

161
P. ﬂu. 33512
P. ﬂu. F 17513

§§§

123

1:3

'- 8'- P

m "N <0--'0'”ooooooc N
"" . .§§€. 3°

ﬁwséiéﬁéﬁéﬁﬁii

1QRRRRRQRRRRRRRRRRRRRRKQRS

 

 

mewmmm

18:010Me 17:0201-1 17:030H 17:01.0 19:01eo 19:Oer1te 19:0cycm8c 18:1201-1

 

 

Table A1. (continued)

Q4

19:0 10 Me 20:3 «6.9.120 20:2 «16.90

12
15
11
19
11
5
76
6
6
1

13

§egzgaazgzgsggglgﬁs131°1111111111123:
3 g 0 N5 8' éBﬁ'
" "' ""‘ E 333332“
'- P" '- X"
5 N kg 38 : E=EE3_8§~~~~~~ at?
o§§93§§°°2 “$38:§§<0¢¢30”'E§§§§§§§ '2
3': as ' 33 g a:
g ‘3‘ 1': gamma 3%
- IL ‘ " '-- ...... .--
l 1 ‘1 “inadmmmmmmll

ERR888RRREa$888RRRRRRRﬁRRﬁRRRRRRRQﬂﬁS

 

125

 

..N
min
a. 5
QVN
adn
ado
0.0?
0.9
v.8
06v
adv
ﬁmn
cfn
oNn
..Nn
..N»
v. ..n
QNN
ndN
v.mN
n. .N
v.0.
o. .N
Nu.
m. .N
nNN
CNN
QVN
hm.
QNN
Nd.
EMN
QNN
N. .N
N.0N
a. .N

5&5 0&5 0‘3 :5

 

1010
INN-Iltllll

N.
0

—.mp
m. 3
o. 3
N. .—
adv
m. 3
Q:
0.9
Na
NN—
0.0.
v.9
of.
1.6.

V.
p

1.1.
06
m6 I
I md
ad Nd
No of
I I 0.0
I I v6

N.
o

888. no. 81

Qm
Ni
0.0
0.0

95

 

2
.88 551.1
SR. 831.1
.9... 3.23.1.1
.21 .882.- .1
892.. 23.58 .1
892.. Sean .1
892.. Eve» .1
89:. 2.5:.» .1
082.. 2.58 .1
8.2.. 2.5:.» .1
.81.. 831.1
8.8 531.1
80.. o .8 .1
3a .8 .1
1.2.. .3. .1
92.. .5. .1
1.32m 5.. .1
«81.. < .8 .1
8.
.N.
n:
«a
x.

B
98.1 .3 .1
~58 5.. .1
.o.

.81.. o .8 .1
8.4.. o .3 .1
8.

8.
no.

.2.

8.

8.
.858 .38.

§man:2.amamamaximumaaaaaauaana'aaaaas

.-

28:528. .2 23

126

 

a 01:9 mm”: x:

on! _.

I °.N II II II II. II
II N. F I II II I II
1' N.‘ I .l | II I
II o.~ ' II II I II

019?. IOmouN— 809.3 :8 —N— :80an sac-OHM. 0190p

 

252.5 32.81

mm I NN
I I n.—
o.. I I
ﬂ. 1 I
N; I I
ad I I
0.. I I

9N— NIme x: 1866—

 

8.

m.o.. 202 8. ca.

m5: 902 8. .92

m5: 202 8. a?
.5 .s
.E .s

8888

m.o.. 202

ms: 202

m5: 902 8. .21

m5: 202 8. ca

«.8. £12280 .1

28. 812228 .1
....

«.8. m .2. .1
289 .5... .8. .02
2.20. .28 8. a?
3.

.a.

8
8
8
a.

3:5: 010‘. coxah

NNN
553330100

888888888

82.28. .2 as...

Table A1. (continued)

 

 

 

Fatty acid (percent)

15211006 15:10010A 111114.503 1520100 15200010 15:1w60

0.3

16:1m7celc 16:1hoH 16:0Ne1c 16:01e018:1oo11c

150

127

16.4

AA oﬁmmmmmmm
2 §§N §£§§§§§§§
0v; FFFFFFF Pv-
~.""'.f~ m
§£SSS§§§§3§§E§§§§§§§§
3 §§5 §§§§§§§§§
3* téﬁﬁsssss
gassasssssasassasgssa

 

128

 

ID
llllolllllllllllllll

.33 (083:.— 0290.2 8885?:

0.
...

8800. EL.

 

888. no. 81

0.
P

008 01 p”:

01,
n

:00 0.. cum.

QMN
odN
odN
mdn
... ..n
0.00
N.Nn
N...“
000
min
v. .0
Eva
1N0
n. 5
him
v.8
9mm
v.00
m.nn
QB
92

m. . ...« ..
1 a. m o.
.. 1..» 1...
.. «.u «.«
.. .8 o.
1 o8 ..«
.. as «.«
1 ..o. n.
.. o.«. 1
II °.m II
1 mm. 1
II F.8 II

no ..m« 1
l “.mﬂ II

no o. .« q.
1 3. o.
.. «.8 .3
II N.& In
.. 2.. a...
.. ..t 3

owe to— one how 88 to.

 

8.
m5: 902.
m5: 902.
m5: 202.
m5: 902.
m5: 902.
m5: 902.
m5: 902
«.8. £122
98. £122
....
«.8. m .2. .1

280. .28 .E .3.
2.20. .28 .E a?

8.

.m.

8

8

8

8

3:5: 380..

€§EE§EE§EE§E
qlﬁiﬁﬁﬁﬁﬁ

282.88. ._< 2.28

8888888918

P‘-
I'D")

g888888888

...

 

129

1 n. 1 1 1 1 1 1 ...« 1 no 1 1 8. 8

 

n. o... 1 1 1 1 1 1 98 1 1 1 1 m5: 902 .E a? «n
«.. m8 1 1 1 1 1 1 no. 1 1 1 1 82. 902 .E .3. «n
«.. 8 1 1 1 1 1 1 o..« 1 1 1 1 ms: 202 E .3. «n
... «.8 1 1 1 1 1 1 no. 1 1 1 1 82.902 E .3. «n
... m.« 1 1 1 1 1 1 98 1 1 1 1 8a.. 902 .E o? 8
1 ...« 1 1 1 1 1 1 m8 1 1 1 1 m5: 902 .E .3. «n
3 o. 1 1 1 1 1 1 «.«. 1 1 1 1 82. 902 .E .2... «n
1 1 1 1 1 1 1 1 no. 1 1 1 1 «.8. £128 2&8 .1 .n
1 1 1 1 1 1 1 1 o.v« 1 1 1 1 28. £122 228 .1 .n
1 1 1 1 1 1 1 1 as. 1 1 1 1 .... .n
1 1 1 1 1 1 1 1 1.... 1 1 1 1 «.8. m .8 .1 8
1 n. 1 1 1 1 1 1 m.«. 1 1 1 1 289 .18 .E .5. 8
1 o. 1 1 1 1 1 1 m.«. 1 1 1 1 25.0. .28 .E .02 8
1 o. 1 no 1 1 1 1 1...... 1 1 1 1 8. 8
no 8 1 «.o 1 1 1 1 o.«. 1 1 1 1 a. 8
1 q. 1 no 1 1 1 1 Q». 1 1 1 1 8 8
1 a. 1 1 1 1 1 1 no. 1 1 1 1 8 8
1 a. 1 no 1 1 1 1 ..o. 1 1 1 1 . 8 8
1 m.« 1 o. 1 1 1 1 3. 1 1 1 1 8 8
89 .6. 8s .5. 8.98 :8 ca. 08. .5. :8 3. 288.98 9.. 263.. 83.1.. 89.1.. or. 3.. 831.. 858.18. :98
88.8. 2881

 

 

98:88. .2 2%..

130

 

868 «now outmda ﬂow 0.2 o. 06.

ION .6.

Na I I
ad I I
0N I I
mN I I
0.. I I
..N I I
NN I I
s... I n6
one go 06.. 8:0 99 on. one

 

Aggy Ho- 33...

md 1 1 1
8.3.. .89.. :83. 2228.

 

one.

8.
m5: 202 .E .0?
m2... 202 E .3
m5: 202 .E a?
m.o.. 202 .E 5?
m.o.. 902 .E .02
m5: 902 .E .3.
m5: 202 .E .81
«.8. 812228 .1
0.8.. £12225 .1
....

«.2. m s. .1
89 :8 .E .02
:29 .28 .E a?

3.
.a.
8
8
8
2.
.0053: 80.02 :0qu

888888888

v-v-
(06')

888888888

8:538. .11 2%.

131

 

Nap
v.2
8m-
mém
0.0..
NN.
m8.
as
5.2
0.2.
ad.
ad
v. . F
o. — —
0d

:5 0&5 £5 :5 an?)

 

I. N." l
1 1 a...
1 1 ...m
1 1 mm

.88. no. .81

 

8.
9a.. 902.
m8: 202
ms: 902.
m5: 902.
m5: 202.
m5: 202.
m6: 202
«.8. £82238 .1
22.. «212225 .1

v..

«.8. m .8 .1
68. .c8 .E .3
:29 88 .E .31

3.

a.

8

8

8

2.
3053:0102 :98...

EEEEEEE
88888888

.0?
.0?
.0?
o?
.03.
.9?
6?

v-v-v-
(01'3")

888888888

 

8:88. .2 03.;

LIST OF REFERENCES

LIST OF REFERENCES

Abou Seada, M. N. 1., and J. C. G. Ottow. 1985. Effect of increasing oxygen
concentration on total denitriﬁcation and nitrous oxide release from soil by different
bacteria. Biology and Fertility of Soils 1:31-38.

Allen, B. B., M. F. Allen, D. J. Helm, J. M. Trappe, R. Molina, E. Rincon. 1995.
Patterns and regulation of mycorrhizal plant and fungal diversity. Pages 47-62 in H.
P. Collins, G. P. Robertson, and M. J. Klug, editors. The Signiﬁcance and Regulation
of Soil Biodiversity. Kluwer Academic Publishers, Dordrecht, The Netherlands.

Alexander, M. 1985. Ecological constraints on nitrogen ﬁxation in agricultural
ecosystems. Advances in Microbial Ecology 8:163-183.

Ambus, P. 1996. Production of N20 in soil during decomposition of dead yeast cells
with different spatial distributions. Plant and Soil 181 :7-12.

Anderson, I. C. and J. S. Levine. 1986. Relative rates of nitric oxide and nitrous oxide
production by nitriﬁers, denitriﬁers, and nitrate respirers. Applied and Environmental
Microbiology 51. 938- 945.

Applied Biostatistics, Inc. 1992. NTSYS. Exeter Publishers, Setauket New York, USA.

Arah, J. R. M., and K. A. Smith. 1989. Steady-state denitriﬁcation in aggregated soils: a
mathematical model. Journal of Soil Science 40:139-149.

Arah, J. R. M., and K. A. Smith. 1990. Factors inﬂuencing the fraction of the gaseous
products of soil denitriﬁcation evolved to the atmosphere as nitrous oxide. Pages
475-480 in A. F. Bouwman, editor. Soils and the Greenhouse Effect. John Wiley and
Sons Ltd., Chichester.

Balderston, W. L., B. Sherr, and W. J. Payne. 1976. Blockage by acetylene of nitrous
oxide reduction in Psedomonas perfectomarinus. Applied and Environmental
Microbiology 31:504-508.

Beare, M. H., D. C. Coleman, D. A. Crossley, P. F. Hendrix, E. P. Odum. 1995. A
hierarchical approach to evaluating the signiﬁcance of soil biodiversity to
biogeochemical cycling. Pages 5-22 in H. P. Collins, G. P. Robertson, and M. J. Klug,
editors. The Signiﬁcance and Regulation of Soil Biodiversity. Kluwer Academic
Publishers, Dordrecht, The Netherlands.

Betlach, M. R., and J. M. Tiedje. 1981. Kinetic explanation for accumulation of nitrite,

nitric oxide, and nitrous oxide during bacterial denitriﬁcation. Applied and
Environmental Microbiology 42: 1074-1084.

132

133

Berg, P., Klemmedtsson, and T. Roswall. 1982. Inhibitory effect of low partial pressures
of acetylene on nitriﬁcation. Soil Biology and Biochemistry 14:301-303.

Binnerup, S. J ., and J. Sorenson. 1992. Nitrate and nitrite microgradients in barley
rhizosphere as detected by a highly sensitive denitriﬁcation bioassay. Applied and
Environmental Microbiology 58:2375-2380.

Blackmer, AM. and J .M. Bremner. 1978. Inhibitory effect of nitrate on reduction of
N20 to N2 by soil microorganisms. Soil Biology and Biochemistry 10:187-191.
Blackmer, AM. and J .M. Bremner. 1979. Stimulatory effect of nitrate on reduction of
N20 to N2 by soil microorganisms. Soil Biology and Biochemistry 11:313-315.

Bobbie, R. J. and D. C. White. 1980. Characterization of benthic microbial community
structure by high-resolution gas chromatography of fatty acid methyl esters. Applied
and Environmental Microbiology 39: 1212- 1222.

Bolin, B. 1998. The keys to negotiations on climate change: a science perspective.
Science 279:330-331.

Bremner, J. M., and A. M. Blackmer. 1978. Nitrous oxide emission from soils during
nitriﬁcation of fertilizer nitrogen. Science 199:295-296.

Brussaard, L., L. Bouwman, M. Geurs, J. Hassink, and K. B. Zwart. 1990. Biomass,
composition and temporal dynamics of soil organisms of a silt loam soil under
conventional and integrated management. Netherlands Journal of Agricultural
Science 38:283-302.

Burth, 1., and J. C. G. Ottow. 1983. Inﬂuence of pH on the production of N20 and N2 by
different denitrifying bacteria and F usarium solani. Ecological Bulletin 35:207-215.

Carlson, C. A., and J. L. Ingraham. 1983. Comparison of denitriﬁcation by
Pseudomonas stutzeri, Pseudomonas aeruginosa, and Paracoccus denitriﬁcans.
Applied and Environmental Microbiology 45:1247-1253.

Cavigelli, M. A., G. P. Robertson, and M. J. Klug. 1995. Fatty acid methyl ester
(FAME) proﬁles as measures of soil microbial community structure. Plant and Soil
170:99-113.

Chapin, F .S., 111, CE. Sala, I.C. Burke, J .P. Grime, D.U. Hooper, W.K. Lauenroth, A.
Lombard, H.A. Mooney, A.R. Mosier, S. Naeem, S.W. Pacala, J. Roy, W. L. Steffen
and D. Tilrnan. 1998. Ecosystem consequences of changing biodiversity:
experimental evidence and a research agenda for the future. Bioscience 48:45-52

Christensen, S., and G. J. Bonde. 1985. Seasonal variation in numbers and activity of
denitrifying bacteria and physiological groups among isolates. Planteavl. 89:367-
372.

Christensen, S., S. Simkins and J.M. Tiedje. 1990a. Spatial variation in denitriﬁcation:
dependency of activity centers on the soil environment. Soil Science Society of
America Journal 54: 1608- 1 61 3.

134

Christensen, 8., S. Simkins and J .M. Tiedje. 1990b. Temporal patterns of soil
denitriﬁcation: their stability and causes. Soil Science Society of America Journal
54:1614-1618.

Christensen, 8., and J. M. Tiedje. 1988. Sub-parts-per—billion nitrate method. Use of an
N20- producing denitriﬁer to convert N03 or NO; to N20. Applied and
Environmental Microbiolog 54: 1409- 1413.

Christensen, S., and J. M. Tiedje. 1990. Brief and vigorous N20 production by soil at
sping thaw. Journal of Soil Science 41 :1-4.

Cicerone, R. J. 1987. Changes in stratospheric ozone. Science 237:35-41.

Collins, H. P., G. P. Robertson, M. J. Rosek, S. A. Gage, and J. R. Crum. 1998. Soils
and land use patterns within a fragmented landscape of southwest Michigan. Soil
Survey Horizons (in press).

Conrad, R. 1996. Soil microorganisms as controllers of atmospheric trace gases (H2,
CO, CH4, OCS, N20 and NO). Microbiological Reviews 60:609-640.

Costilow, RN. 1981. Biophysical factors in growth. Pages 66-78 in P. Gerhardt, editor.
Manual of Methods for General Bacteriology. American Society for Microbiology,
Washington, DC.

Davidson, E. A. 1991. Fluxes of nitrous oxide and nitric oxide from terrestrial
ecosystems. Pages 219-235 in J. E. Rogers and W. B. Whitman, editors. Microbial
Production and Consumption of Greenhouse Gases. American Society for
Microbiology, Washington, DC, USA.

Davidson, B. A., W. T. Swank, and T. 0. Perry. 1986. Distinguishing between
nitriﬁcation and denitriﬁcation as sources of gaseous nitrogen production in soil.
Applied and Environmental Microbiology 52: 1280-1286.

Dendooven, L. and J. M. Anderson. 1994. Dynamics of reduction enzymes involved in
the denitriﬁcation process in pasture soil. Soil Biology and Biochemistry 26: 1501-
1506.

Dendooven, L., E. Pemberton, and J. M. Anderson. 1996. Denitriﬁcation potential and
reduction enzyme dynamics in a Norway spruce plantation. Soil Biology and
Biochemistry 28: 15 1-157.

Di gby, P. G. N. and R. A. Kempton. 1987. Multivariate Analysis of Ecological
Communities. Chapman and Hall, New York, USA.

Dunning, J. B., B. J. Danielson, and H. R. Pulliarn. 1992. Ecological processes that
affect populations in complex landscapes. Oikos 65:169-175.

Erwin, J. A. 1973. Fatty acids in eukaryotic microorganisms. Pages 41-143 in J. A.
Erwin, editor. Lipids and Biomembranes of Eukaryotic Microorganisms. Academic
Press, New York, USA.

135

Everitt, B. 1980. Cluster Analysis. Halsted Press, New York, USA.

Findlay, R. H. and D. C. White. 1983. The effects of feeding by the sand dollar Mellita
quinquiesperforata (Leske) on the benthic microbial community. Journal of
Experimental Marine Biology and Ecology 72:25-41 .

Firestone, M. K., and E. A. Davidson. 1989. Microbiological basis of NO and N20
production and consumption in soil. Pages 7-21 in M. O. Andreae and D. S. Schimel,
editors. Exchange of Trace Gases Between Terrestrial Ecosystems and the
Atmosphere. John Wiley & Sons Ltd., Berlin.

Firestone, M. K., R. B. Firestone, and J. M. Tiedje. 1980. Nitrous oxide from soil
denitriﬁcation: factors controlling its biological production. Science 208:749-751.

Firestone, M. K., M. S. Smith, R. B. Firestone, and J. M. Tiedje. 1979. The inﬂuence of
nitrate, nitrite and oxygen on the composition of gaseous products of denitriﬁcation in
soil. Soil Science Society of America Journal 43:1 140-1144.

Firestone, M. K., and J. M. Tiedje. 1979. Temporal change in N20 and N2 from
denitriﬁcation following onset of anaerobiosis. Applied and Environmental
Microbiology 38:673-679.

Focht, P. D. 1974. The effect of temperature, pH and aeration on the production of
nitrous oxide and gaseous nitrogen -- a zero-order kinetic model. Soil Science
118:173-179.

Foster, R. C. 1988. Microenvironments of soil microorganisms. Biology and Fertility of
Soils 62189-203.

Fries, M. R., J. Zhou, J. Chee-Sanford, and J. M. Tiedje. 1994. Isolation,
characterization, and distribution of denitrifying toluene degraders from a variety of
habitats. Applied and Environmental Microbiology 60:2802-2810.

Frostegaard, A. and E. Bath. 1996. the use of phospholipid fatty acid analysis to
estimate bacterial and fungal biomass in soil. Biology and Fertility of Soils 22:59-65.

Gamble, T. N., M. R. Betlach, and J. M. Tiedje. 1977. Numerically dominant
denitrifying bacteria from world soils. Applied and Environmental Microbiology
33:926-939.

Gillan, F. T. and R. W. Hogg. 1984. A method for the estimation of bacterial biomass
and community structure in mangrove-associated sediments. Journal of
Microbiological Methods 22275-293.

Goreau, T.J., W.A. Kaplan, S.C. Wofsy, M.B. McElroy, F.W. Valois, and SW. Watson.
1980. Production of N02' and N20 by nitrifying bacteria at reduced concentrations of
oxygen. Applied and Environmental Microbiology 40:526-532.

Gorlach, K., R. Shingaki, H. Morisaki, and T. Hattori. 1994. Construction of eco-
collection of paddy ﬁeld soil bacteria for population analysis. Journal of General and
Applied Microbiology 40:509-517.

136

Grime, J .P. 1997. Biodiversity and ecosystem ﬁrnction: the debate deepens. Science
277: 1260-1261.

Groffman, P. M. and J. M. Tiedje. 1989. Denitriﬁcation in north temperate forest soils:
spatial and temporal patterns at the landscape and seasonal scales. Soil Biology and
Biochemistry 21:613-620.

Harwood, J. L. and N. J. Russell. 1984. Lipids in Plants and Microbes. Allen & Unwin,
London.

Hattori, T. 1988. Soil aggregates as microhabitats of microorganisms. Report of the
Institute for Agricultural Research T ohoku University. 37:23-36.

Heal, O. W., S. Struwe, and A. Kjoller. 1996. Diversity of soil biota and ecosystem
function. Pages 385-402 in B. Walker and W. Steffen, editors. Global Change and
Terrestrial Ecosystems. Cambridge University Press, Cambridge, England.

Hinkelman, K. and O. Kempthome. 1994. Design and Analysis of Experiments, volume
1. Wiley and Sons, New York, USA.

Hochstein, L. 1., and G. A. Tomlinson. 1988. The enzymes associated with
denitriﬁcation. Annual Review of Microbiolog» 42:231-262.

Hojber, 0., H. S. Johansen, and J. Sorensen. 1994. Determination of 15N abundance in
nanograrn pools of NO3' and N02' by denitriﬁcation bioassay and mass spectrometry.
Applied and Environmental Microbiology 60:2467-2472.

Hooper, D. U. and P. M. Vitousek. 1997. The effects of plant composition and diversity
on ecosystem processes. Science 277: 1302-1305.

Houghton, J .T., L. G. Meira F ilho, B. A. Callander, N. Harris, A. Kattenberg, and K.
Maskell, editors. 1996. Climate Change 1995: The Science of Climate Change.
Cambridge University Press, Cambridge, England.

Hunter, M. D., T. Ohgushi, and P. W. Price. 1992. Effects of Resource Distribution on
Animal-Plant Interactions. Academic Press, San Diego.

Huston, M. 1997. Hidden treatments in ecological experiments: re-evaluating the
ecosystem function of biodiversity. 0ecologia 110: 449-460.

Jantzen, E. and K. Bryn. 1985. Whole-cell and lipopolysaccharide fatty acids and sugars
of Gram-negative bacteria. Pages 145-171 in M. Goodfellow and D. E. Minnikin,
editors. Chemical Methods in Bacterial Systematics. Academic Press, London.

J oliffe, I. T. 1986. Principal Component Analysis. Springer Verlag, Berlin, Germany.
Johnsson, H., L. Bergstrom, P.-E. Jansson, and K. Paustian. 1987. Simulation of

nitrogen dynamics and losses in a layered agricultural soil. Agriculture Ecosystems
and Environment 18:333-356.

137

Ka, J .-O., J. Urbance, R. W. Ye, T.-Y. Ahn, and J. M. Tiedje. 1997. Diversity of oxygen
and N-oxide regulation of nitrite reductase in denitrifying bacteria. FEMS
Microbiology Letter 156255-60.

Kadmon, R. 1993. Population dynamic consequences of habitat heterogeneity: an
experimental study. Ecology 74:816-825.

Klemmedtsson, L., B. H. Svensson, and T. Rosswall. 1988a. A method of selective
inhibition to distinguish between nitriﬁcation and denitriﬁcation as sources of N20 in
soil. Biology and Fertility of Soils 6:1 12-119.

Klemmedtsson, L., B. H. Svensson and T. Rosswall. 1988b. Relationships between soil
moisture content and nitrous oxide production during nitriﬁcation and denitriﬁcation.
Biology and Fertility of Soils 6:106-111.

Klug, M.J. and J .M. Tiedje. 1994. Response of microbial communities to changing
environmental conditions: chemical and physiological approaches. Pages 371-378 in
R. Guerrero and C. Pedros-Alio, editors. Trends in Microbial Ecology. Spanish
Society for Microbiology, Barcelona, Spain. .

Komer, H. 1993. An aerobic expression of nitric oxide reductase from denitrifying
Pseudomonas stutzeri. Archives of Microbiology 1592410-416.

Koskinen, W. C. and D. R. Keeney. 1982. Effect of pH on the rate of gaseous products
of denitriﬁcation in a silt loam soil. Soil Science Society of America Journal 46:1165-
1167.

Lensi, R., F. Gourbiere, and A. Josserand. 1985. Measurement of small amounts of
nitrate in an acid soil by N20 production. Soil Biology and Biochemistry 17:733-734.

Li, C., S. Frolking, and T. A. Frolking. 1992a. A model of nitrous oxide evolution from
soil driven by rainfall events. 1. Model structure and sensitivity. Journal of
Geophysical Research 97:9759-9776.

Li, C., S. Frolking, and T. A. Frolking. 1992b. A model of nitrous oxide evolution from
soil driven by rainfall events. 2. Model applications. Journal of Geophysical Research
97:9777-9783.

Liu, C. J. and T. D. Keister. 1978. Southern pine stern form deﬁned through principal
component analysis. Canadian Journal of Forestry Research 8: 188-197.

Lynch, 1. M. 1986. Rhizosphere microbiology and its manipulation. Biological
Agriculture and Horticulture 3: 143-152.

Mahne, I. and J .M. Tiedje. 1995. Criteria and methodology for identifying respiratory
denitriﬁers. Applied and Environmental Microbiology 61:1 110-1115.

Mayberry, W. R., D. W. Lambe, and K. P. Ferguson. 1982. Identiﬁcation of Bacteroides
species by cellular fatty acid proﬁles. International Journal of Systematic
Bacteriology 32:21 -27.

138

McConnaughey, P. K. and D. R. Bouldin. 1985. Transient microsite models of
denitriﬁcation: 1. Model development. Soil Science Society of America Journal
49:886-891.

McGill, W. B., H. W. Hunt, R. G. Woodmansee, and J. O. Reuss. 1981. Pheonix, a
model of the dynamics of carbon and nitrogen in grassland soils. Ecology Bulletin
33249-115.

Meyer, 0. 1993. Functional groups of microorganisms. Pages 67-96 in E.-D. Schulze
and H. A. Mooney, editors. Biodiversity and Ecosystem Function. Springer-Verlag,
Berlin, Germany.

Microbial ID Inc. 1992. Microbial Identiﬁcation System Operating Manual, Version 4.
Newark, Delaware, USA.

Milligan, G. W., and M. C. Cooper. 1987. Methodological reviews: clustering methods.
Applied Psychological Measurement 1 1:329-354.

Mooney, H. A., J. Lubchenco, R. Dirzo, and O. E. Sala. 1995. Biodiversity and
ecosystem functioning: basic principles. Pages 289-301 in V. H. Heywood, editor.
Global Biodiversity Assessment. Cambridge University Press, Cambridge, UK.

Moss, C. W. 1981. Gas-liquid chromatography as an analytical tool in microbiology.
Journal of Chromatography 203:337-347.

Moss, C. W., S. B. Dees, and G. O. Guerrant. 1980. Gas-liquid chromatography of
bacterial fatty acids with a fused-silica capillary column. Journal of Clinical
Microbiology 12:127-130.

Munch, J. C. 1989. Organism speciﬁc denitriﬁcation in samples of an Udiﬂuvent with
different nitrate concentrations. Z. Pﬂanzenernahr. Bodenk 152:395-400.

Munch, J. C. 1991. Nitrous oxide emission from soil as determined by the composition
of denitrifying microbial population. Pages 309-316 in , editor. Diversity of
Environmental Biogeochemistry. Elsevier, Amsterdam.

Munch, J. C. and J. C. G. Ottow. 1986. Nature des produits gazeux forrnés dans les sols
a partir de differents microﬂores denitriﬁantes. Science du Sol 24:337-350.

Nommik, 1956. Investigations on denitriﬁcation in soil. Acta Agric. Scand. 6:195-228.

Ojima, D., C. Bledsoe, P. Matson, A. Mosier, J. Melillo, and G. P. Robertson. 1992.
Building a US Trace Gas Network. Report of MAB and IGBP/IGAC workshop held
at Pingree Park, Colorado, September.

Ott, L. 1984. An Introduction to Statistical Methods and Data Analysis, 2nd edition.
Duxbury Press, Boston.

Parkin, T. B. 1987. Soil microsites as a source of denitriﬁcation variability. Soil Science
Society of America Journal 51:1 194-1 199.

139

Parkin, T. B., A. J. Sexstone, and J. M. Tiedje. 1985. Adaptation of denitrifying
populations to low soil pH. Applied and Environmental Microbiology 49:1053-1056.

Parton, W. J ., A. R. Mosier, and D. S. Schimel. 1988. Rates and pathways of nitrous
oxide production in a shortgrass steppe. Biogeochemistry 6:45-58.

Parton, W. J ., A. R. Mosier, D. S. Ojima, D. W. Valentine, D. S. Schimel, K. Weier, and
A. E. Kulmala. 1995. Generalized model for N2 and N20 production from
nitriﬁcation and denitriﬁcation. Global Biogeochemical Cycles 10:401-412.

Perry, G. J ., J. K. Volman, R. B. Johns, and H. J. Bavor. 1979. Fatty acids of bacterial
origin in contemporary marine sediments. Geochimica Cosmochimica Acta 43: 1 715-
1725.

Potvin, C. 1993. ANOVA: experiments in controlled environments. Pages 46-68 in S.
M. Scheiner and J. Gurevitch, editors. Design and Analysis of Ecological
Experiments. Chapman and Hall, New York, USA.

Rasmussen, R. A. and M. A. K. Khalil. 1986. Atmospheric trace gases. trends and
distributions over the last decade. Science 232: 1623- 1624.

Rice, C. W., P. E. Sierzega, J. M. Tiedje, and L. W. Jacobs. 1988. Stimulated
denitriﬁcation in the microenvironment of a biodegradable organic waste injected into
soil. Soil Science Society of America Journal 52:102-108.

Richaume, A., C. Steinbers, L. Jocteur Monrozier, and G. Faurie. 1993. Differences
between direct and indirect enumeration of soil bacteria: the inﬂuence of soil structure
and cell location. Soil Biology and Biochemistry 25:641-643.

Robertson, G. P. 1993. Fluxes of nitrous oxide and other nitrogen gases from intensively
managed landscapes: a global perspective. Pages 95-108 in L. A. Harper, A. R.
Mosier, J. M. Duxbury, and D. E. Rolston, editors. Agricultural Ecosystem Effects
on Trace Gases and Global Climate Change. American Society of Agronomy,
Madison, Wisconsin, USA.

Robertson, G. P. and K. Gross. 1994. Assessing the heterogeneity of belowground
resources: quantifying pattern and scale. Pages 237-253 in M. M. Caldwell and R. W.
Pearcy, editors. Exploitation of Environmental Heterogeneity by Plants. Academic
Press, New York, USA.

Robertson, G. P., K. M. Klingensmith, M. J. Klug, E. A. Paul, J. R. Crum, and B. G.
Ellis. 1997. Soil resources, microbial activity, and primary production across an
agricultural ecosystem. Ecological Applications 7:158-170.

Robertson, L. A., T. Dalsgaard, N. -P. Revsbech, and J. G. Kuenen. 1995. Confirmation
of‘ aerobic denitriﬁcation’ in batch cultures, using gas chromatography and Nmass
spectrometry. FEMS Microbiology Ecology 18: 113- 120.

Rolston, D. E., P. S. C. Rao, J. M. Davidson, and R. E. Jessup. 1984. Simulation of
denitriﬁcation losses of nitrate fertilizer applied to uncropped and manure-amended
ﬁeld plots. Soil Science 137:270-279.

140

Saertre, P. 1998. Decomposition, microbial community structure, and earthworm effects
along a birch-spruce soil gradient. Ecology in press.

SAS Institute. 1991. SAS/STAT User’s Guide. Release 6.03. SAS Institute, Cary,
North Carolina, USA.

SAS Institute. 1996. SAS/STAT User’s Guide. Release 6.09 enhanced. SAS Institute,
Cary, North Carolina, USA.

Sasser, M. 1990. Identiﬁcation of bacteria by gas chromatography of cellular fatty acids.
Technical Note # 101. MIDI Inc., North Newark, DE.

Schimel, J. 1995. Ecosystem consequences of microbial diversity and community
structure. Pages 237-252 in F. S. Chapin HI, and C. Komer, editors. Arctic and Alpine
Biodiversity: Patterns, Causes and Ecosystem Consequences. Springer-Verlag,
Berlin, Germany.

Seech, A. G. and E. G. Beauchamp. 1988. Denitriﬁcation in soil aggregates of different
sizes. Soil Science Society of America Journal 52:1616-1621.

Sexstone, A. J ., T. B. Parkin and J. M. Tiedje. 1985a. Temporal response of soil
denitriﬁcation rates to rainfall and irrigation. Soil Science Society of America Journal
49:99-103.

Sexstone, A. J ., N. P. Revsbech, T. B. Parkin and J. M. Tiedje. 1985b. Direct
measurement of oxygen proﬁles and denitriﬁcation rates in soil aggregates. Soil
Science Society of America Journal 49:645-651.

Shannon, CE. and W. Weaver. 1949. The Mathematical Theory of Communication, p.
117. University of Illinois Press, Urbana, IL.

Smith, K. A., G. P. Robertson, and J. M. Melillo. 1993. Exchange of trace gases
between the terrestrial biosphere and the atmosphere in the midlatitudcs. Pages 179-
203 in R. G. Prinn, editor. Global Atmospheric-Biospheric Chemistry. Plenum Press,
New York, USA.

Smith, M. S., M. K. Firestone, and J. M. Tiedje. .1978. Acetylene inhibitti method for
short-term measurement of soil denitriﬁcation and its evaluation using N. Soil
Science Society of America Journal 42:61 1-615.

Smith, M. S., and L. L. Parsons. 1985. Persistence of denitrifying enzyme activity in
dried soils. Applied and Environmental Microbiology 49:316-320.

Smith, M. S., and J. M. Tiedje. 1979. Phases of denitriﬁcation following oxygen
depletion in soil. Soil Biology and Biochemistry 11:261-267.

Stackebrandt, E. 1992. Unifying phylogeny and phenotypic diversity. Pages 24-33 in A.
Balows, editor. The Prokaryotes: A Handbook on the Biology of Bacteria:
Ecophysiology, Isolation, Identiﬁcation, Applications. Springer-Verlag, New York,
USA.

141

Tiedje, J. M. 1988. Ecology of denitriﬁcation and dissimilatory nitrate reduction to
ammonium. Pages 179-244 in A. J. B. Zehnder, editor. Biology of Anaerobic
Microorganisms. John Wiley & Sons, Chichester, England.

Tiedje, J. M. 1994. Denitriﬁers. Pages 245-267 in R.W. Weaver, J. S. Angle, and P. S.
Bottomley, editors. Methods of Soil Analysis, Part 2. Microbiological and
Biochemical Properties. Soil Science Society of America, Madison, Wisconsin, USA.

Tiedje, J .M., A.J. Sexstone, D.D. Myrold, and J .A. Robinson. 1982. Denitriﬁcation:
ecological niches, competition and survival. Antonie van Leeuwenhoek 48:569-583.

Tilrnan, D., J. Knops, D. Wedin, P. Reich, M. Ritchie, and E. Siemann. 1997. The
inﬂuence of functional diversity and composition on ecosystem processes. Science
277:1300-1302.

Torsvik, V., J. Goksoyr, and F. L. Daae. 1990a. High diversity in DNA of soil bacteria.
Applied and Environmental Microbiology 56:782-787.

Torsvik, V., K. Salte, R. Sorheim, and J. Goksoyr. 1990b. Comparison of phenotypic
diversity and DNA heterogeneity in a population of soil bacteria. Applied and
Environmental Microbiology 56:776-781.

Trangmar, B. B., R. S. Yost, and G. Uehara. 1985. Application of geostatistics to Spatial
studies of soil properties. Advances in Agronomy 38:45-94.

Tsuji, T., Y. Kwasaki, S. Takeshima, T. Sekiya, and S. Tanaka. 1995. A new
ﬂuorescence staining assay for visualizing living microorganisms in soil. Applied
and Environmental Microbiology 61 :3415-3421.

Tumer, M. G. 1989. Landscape ecology: the effect of pattern on process. Annual
Review of Ecology and Systematics. 20:171-197.

Urakami, T., C. Ito-Yoshida, H. Araki, T. Kijima, K.-I. Suzuki and K. Komagata. 1994.
Transfer of Pseudomonas plantarii and Pseudomonas glumae to Burkholderia as
Burkholderia spp. and description of Burkholderia vandii sp. nov. International
Journal of Systematic Bacteriology 44:235-245.

Vestal, J. R. and D. C. White. 1989. Lipid analysis in microbial ecology. Bioscience
39:535-541.

Viljoen, B. C., J. L. F. Kock, and P. M.. Lategan. 1986. Fatty acid composition as a
guide to the classiﬁcation of selected genera of yeasts beleonging to the
endomycetales. Journal of General Microbiology 132:2397-2400.

Volkman, J. K., R. B. Johns, F. T. Gillan, G. J. Perry, and H. J. Bavor. 1980. Microbial
lipids of an intertidal sediment - 1. Fatty acids and hydrocarbons. Geochimica
Cosmochimica Acta 44:1133-1143.

Webster, R. 1985. Quantitative spatial analysis of soil in the ﬁeld. Advances in Soil
Science 3:1-70.

142

Weier, K. L. and J. W. Gilliam. 1986. Effect of acidity on denitriﬁcation and N20
evolution from Atlantic Coastal Plain soils. Soil Science Society of America Journal
50:1202-5.

Weier, K. L., and I. C. MacRae. 1992. Denitrifying bacteria in the proﬁle of a Brigalow
clay soil beneath a permanent pasture and a cultivated crop. Soil Biology and
Biochemistry 24:919-923.

White, D. C. 1983. Analysis of microorganisms in terms of quantity and activity in
natural environments. Pages 37-66 in J. H. Slater, R. Whittenbury, and J. W. T.
Wimpenny, editors. Microbes in Their Natural Environments. Cambridge University
Press, Cambridge, UK.

Yabuuchi, E., Y. Kosako, H. Oyaizu, I. Yano, H. Hotta, Y. Hashimoto, T. Ezaki and M.
Arakawa. 1992. Proposal of Burkholderia gen. nov. and transfer of seven species of
the genus Pseudomonas homology group H to the genus with the type species
Burkholderia cepacia (Palleroni and Holmes 1981) comb. nov. Microbiol. Immunol. 1
36:1251-1275. '

 

Yoshinari, J. T. and R. Knowles. 1976. Acetylene inhibition of nitrous oxide reductase
by denitrifying bacteria. Biochemical and Biophysical Research Communications
69:705-710.

Zelles, L. and Q. Y. Bai. 1993. Fractionation of fatty acids derived from soil lipids by
solid phase extraction and their quantitative analysis by GC-MS. Soil Biology and
Biochemistry 25:495-507.

Zelles, L., Q. Y. Bai and F. Beese. 1992. Signature fatty acids in phospholipids and
lipopolysaccharides as indicators of microbial biomass and community structure in
agricultural soils. Soil Biology and Biochemistry 24:317-323.

Zuberer, D. A. 1994. Recovery and enumeration of viable bacteria. Pages 245-267 in
R.W. Weaver, J. S. Angle, and P. S. Bottomley, editors. Methods of Soil Analysis,
Part 2. Microbiological and Biochemical Properties. Soil Science Society of
America, Madison, Wisconsin, USA.

Zumft, W. G. 1992. The denitrifying prokaryotes. Pages 554-582 in A. Balows, editor.
The Prokaryotes: A Handbook on the Biology of Bacteria: Ecophysiology, Isolation,
Identiﬁcation, Applications. Springer-Verlag, New York, USA.

Zumﬁ, W. G. 1997. Cell biology and molecular basis of denitriﬁcation. Microbiology
and Molecular Biology Reviews 61:533-616.