v _ .< -
Luann .ngnmmw. .. .
6mm. .2 1qu
. 1. F.

 

._n.. ,..,3 .
.3 $.an

.~
1 3
. . S
«Hymn-a. . ,
. 1. . “4&5!

 

 

1....vl (351“
53.1 .3
a .x
L11

 

.... . 3. 4.. .J.

. '4 éafiwméafim

        

           

 
  
      

 

 

 

 

an. .15}? 15.9.. . . Dr 2 .. t, . .
A: .i L. Lee. Ann: “Levitt 4

mam LIPRARY
J ' Michigan State
, ,m University

This is to certify that the
dissertation entitled

DISTRIBUTION, ABUNDANCE, AND ACTIVITY OF
ANAEROBIC AMMONIUM OXIDIZING BACTERIA AND
MICROBIAL COMMUNITY STRUCTURE IN MARINE
SEDIMENTS

presented by
CHRISTOPHER RYAN PENTON

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Cmp and Soil Sciences

 

 

C

  

 

/ Major Professor’s Sigflature

1 2/09/2008

 

Date

MSU is an Afflnnative Action/Equal Opportunity Employer

 

n-nu-.-.-.-.-.-----.--.---—-—--.-.-.--.—.--.-.-.-.-.—.—.-.—-—-—.-.-v-.-.-—.—-— -

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KrlProi/AccllPresIClRC/DateDue indd

DISTRIBUTION, ABUNDANCE, AND ACTIVITY OF ANAEROBIC AMMONIUM
OXIDIZING BACTERIA AND MICROBIAL COMMUNITY STRUCTURE IN
MARINE SEDIMENTS
By

Christopher Ryan Penton

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Crop and Soil Sciences

2008

ABSTRACT

DISTRIBUTION, ABUNDANCE, AND ACTIVITY OF ANAEROBIC AMMONIUM
OXIDIZING BACTERIA AND MICROBIAL COMMUNITY STRUCTURE IN
MARINE SEDIMENTS
By

Christopher Ryan Penton

Anaerobic ammonium oxidation is a recently discovered nitrogen removal
pathway in natural systems mediated by deep-branching members of the phylum
Planctomycetes, the anammox bacteria. First found to be a significant nitrogen sink in the

Black Sea anoxic water column in 2003, anammox has shown to be responsible for as

much as 79% of N2 production in marine sediments. Therefore, understanding the

distribution, abundance, and activity of anammox communities is essential for refining

estimates of N2 removal in the marine environment. The distribution of the Scalindua sp.

type anammox bacteria was studied in a variety of biogeochemical conditions and diverse
environments, using newly designed 16S rRNA gene primers specific for their target.
Since previous anammox targeted surveys used broad Planctomycetes primers, this was
the first specific screening tool for the identification of anammox bacteria in the
environment that was not based on the use of fluorescence in-situ hybridization. This
primer set was 100% specific in the environments tested for anammox bacteria and
identified the bacteria in a wide variety of environments, ranging from Siberian

permafrost to shallow wetland sediments. Anammox importance in marine sediments has

been thought to increase with depth below water surface, such that the process dominates
over denitn'fication, a result of lower reactive carbon availability. The activity and

abundance of anammox bacteria was studied in the deep marine sediments of Cascadia

Basin using the 15N isotope pairing method and newly designed quantitative PCR

primers, targeting the Scalindua sp. 16S rRNA gene. Anammox was found to contribute

between 12 and 51% of total N2 production in these sediments, far below the dominance

that was expected based on earlier studies. The vertical distribution of anammox
indicated that the bacteria were present in significant numbers at depths where oxidized

N species were not found. This seemed to be indicative of bioturbation creating

preferential porewater channels for N03- / NOz' inputs, transient pulses resulting from

pelagic inputs, or alternative metabolic pathways in Scalindua sp. that does not depend

on oxidized N species. Estimated anammox abundances were highly correlated with

anammox N2 production (R2=O.93) in these sediments, linking the catalyst with the

process. Marine microbial communities in ten diverse geographical and biogeochemical
sediments were studied using 16S rRNA gene targeted pyrosequencing. Geographical
location, temperature, or basin did not influence the relationship among sites. There were
no cosmopolitan clusters, (i.e. sequences grouped at 5% identity), found at every site.
Potentially new groups of bacteria were identified that exhibited less than 85% similarity
to the database, one of which comprised 25% of the community in an Arctic Barrow
Canyon sediment. Rare members were more divergent from the public database,
suggesting that they were not transients but rather localized divergent members that may

play a minor or no role in the functional capacity of the sediment.

ACKNOWLEDGEMENTS

There are many people to whom I wish to convey thanks and admiration towards
in their myriad of roles in shaping not only this dissertation but my academic career. First
I would like to thank Dr. James Tiedje for providing the opportunity to join his lab, for
his inspiration and guidance, for allowing freedom to pursue my interests, and for his
support throughout my career at Michigan State University. I look forward to continue
working with him and members of the Center for Microbial Ecology through my post-
doctoral project. I sincerely thank my other committee members Dr. Terrence Marsh, Dr.
Phil Hamilton, and Dr. Steve Hamilton for their support, their criticisms, and the
opportunity to discuss issues throughout the course of this dissertation. I would also like
to thank Dr. Allan Devol for his collaboration, for allowing me to participate in the
sampling cruise, and for our discussions in the early development of this project. Thanks
to Dr. Pia Engstriim with whom I have had the pleasure of continuing this work with in
Hawaii and Sweden. She introduced me to the applied science of anammox, was
responsible for the chemical and isotope analysis in this dissertation, and was always not
only a colleague but a friend who always laughed, even in the four degree walk in cooler
on a rocking ship. Thanks to Dr. Sue Newman for her encouragement through my very
early days as a graduate student, for laying the foundation of my current career, for her

continued support and making me question everything.

Thanks to all the present and past Tiedje lab members who provided a stimulating
environment that supported open discussions, troubleshooting, and constant feedback on

projects throughout my Ph.D. Many thanks go out to Woo Jun Sul, a cohort through this

time, for the tireless discussions, intriguing conference journeys, and never-ending
analysis of pyrosequencing data. I would also like to apologize for those countless times
my desk chair hit his. I would also like to acknowledge Erick Cardenas who has also
been there throughout my stay at Michigan State for his hallway discussions and talks
about life. Additional thanks goes out to Veronica Gruntzig, Peter Bergholz, Dieter
Tourlousse, Deborah Rodriguez, Stephan Gantner, Deborah Himes, Brian Campbell,
Patrick Chain, Shoko Iwai, and Yu Yang who have all helped in one way or another.
Thanks to the members of the RDP, especially Qiong Wang and Jim Cole, for their
discussions and analysis of pyrosequencing data and for those at the RTSF, especially
Shari Tjugum-Holland and Jeff Landgraf for sequencing clone libraries and for
quantitative PCR.

Besides the people that have contributed directly to my research, there are many
people through the years that have helped me reach this point through their support and
encouragement. Thank you to my family, to my mom and dad who constantly supported
and loved me, who provided those early chemistry sets, who endured scrapes and bruises
as I discovered physics, who nurtured me and showed me a world full of possibilities,
who made sure my homework was done, who helped me at the last minute on those early
science experiments, and who encouraged me through life to be my best.

And now, last but not least, I would like to thank my wife Stephanie who endured
through sleepless nights, who was left at home so many times on my travels, who listened
to presentations, proofread papers, and loved me unconditionally. Thank you for always
being there while we looked for the light at the end of the tunnel in this adventure of ours.

We have finally made it! You are and always have been my angel.

PREFACE

The work presented in this dissertation was part of a collaborative effort on the
behalf of Michigan State University and the University of Washington and funded by the
Department of Energy. The Arctic and Puget Sound samples analyzed were retrieved by
Allan Devol, University of Washington. The GC/MS analyses, porewater calculations,
and flux calculations in chapter three were performed by Pia Engstrdm, University of
Washington, now at Kristenberg Marine Station, University of Gothenburg, Sweden.
Woo Jun Sul of Michigan State University was responsible for the design and

implementation of the pyrosequencing primers used in this study.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................................................................. x
LIST OF FIGURES ............................................................................ xi
CHAPTER 1

INTRODUCTION ............................................................................... 1

OBJECTIVES .................................................................................. 12
Objective 1 .................................................................................. 12
Objective 2 .................................................................................. 14
Objective 3 .................................................................................. 15
Objective 4 .................................................................................. 16

REFERENCES ................................................................................ 17

CHAPTER 2

MOLECULAR EVIDENCE FOR THE BROAD DISTRIBUTION
OF ANAMMOX IN FRESHWATER AND MARINE SEDIMENTS ................. 21

ABSTRACT ...................................................................................... 21
INTRODUCTION .................................................................................... 21
MATERIALS AND METHODS ............................................................... 23
Sampling .......................................................................................... 23
Primer design .................................................................................... 24
DNA extraction ................................................................................... 25
RESULTS AND DISCUSSION ................................................................ 26
Acknowledgements ............................................................................... 38
REFERENCES ..................................................................................... 39

CHAPTER 3

QUANTITATIVE PCR AND ACTIVITY OF ANAEROBIC
AMMONIUM-OXIDIZING BACTERIA IN DEEP-PACIFIC

OCEAN SEDIMENTS OF THE CASCADIA BASIN ..................................... 42
ABSTRACT ...................................................................................... 42
INTRODUCTION ................................................................................ 43

vii

MATERIALS AND METHODS .......................................................... 45

Site description and sediment sampling ................................................. 45
Oxygen and porewater profiles ........................................................... 46
15N incubations and analysis .............................................................. 48
DNA extraction ............................................................................. 50
Prirner-probe design and specificity .................................................... 50
Quantitative PCR assays .................................................................. 51
Q-PCR sensitivity, detection limit, and accuracy ..................................... 52
RESULTS ..................................................................................... 52
Porewater 02 and dissolved inorganic nitrogen ........................................ 52
15 N incubation experiments ............................................................... 53
Quantitative PCR assays .................................................................. 56
DISCUSSION ................................................................................ 62
Absolute rates ............................................................................... 70
Significance for the marine N budget .................................................... 71
Quantitative PCR ........................................................................... 71
Acknowledgements ........................................................................ 75
REFERENCES ................................................................................ 77
CHAPTER 4

COMPARISON OF MICROBIAL COMMUNITY STRUCTURE
AMONG DIVERSE MARINE SEDIMENTS USING

PYROSEQUENCING ........................................................................ 82
ABSTRACT .................................................................................... 82
INTRODUCTION ............................................................................. 83
MATERIALS AND METHODS ........................................................... 85

Sediment sampling .......................................................................... 85
SSU rRNA amplicon pyrosequencing .................................................... 87
Pyrosequencing analysis ................................................................... 88
RESULTS ...................................................................................... 89
Phylogenetic structure ...................................................................... 89
Diversity characterization .................................................................. 92
DISCUSSION ................................................................................ 101
Community diversity ...................................................................... 101
The rare biosphere ......................................................................... 102
Selective effects in Florida Bay .......................................................... 103
Acknowledgements ........................................................................ 106
REFERENCES ............................................................................... 107

viii

CHAPTER 5

QUANTITATIVE PCR OF FRESHWATER ANAMMOX IN

HAWAIIAN TARO FIELDS ............................................................ 110
ABSTRACT ................................................................................ 110
INTRODUCTION .......................................................................... 1 1 1
METHODS .................................................................................. 113

Quantitative PCR assays ................................................................. 113
Sample description and DNA extraction .............................................. 114
RESULTS AND DISCUSSION .......................................................... 114
Acknowledgements ...................................................................... 123
REFERENCES ............................................................................. 124

CHAPTER 6

CONCLUSIONS AND FUTURE PERSPECTIVES ................................ 126
FUTURE PERSPECTIVES ............................................................. .129

APPENDIX

CAN ONICAL CORRESPONDENCE ANALYSIS FOR LINKING
ENVIRONMENTAL VARIABLES WITH MICROBIAL COMMUNITY
STRUCTURE ............................................................................... 132

LIST OF TABLES

TABLE 4.1. Total sequence reads, number of unclassified (Uncl), corresponding
number of clusters (OTUS), Chaol, and Shannon Diversity (H’) values for three
cluster cutoff dissimilarity values ................................................................. 90

TABLE 5.1. Estimated nitrogen removal rates based on quantitative PCR runs
using the 818F/1066R primer set ................................................................ 120

TABLE 5.2. Original and corrected estimated anammox abundances based on
quantitative PCR using the 818F/1066R primer set .......................................... 121

TABLE 5.3. Estimated corrected nitrogen removal rates based on quantitative
PCR runs using the 818F/1066R primer set ................................................... 122

LIST OF FIGURES

FIGURE 1.1. Anaerobic ammonium oxidation pathway of N removal
in context ofthe nitrogen cycleS

FIGURE 1.2. Experimental layout of the isotope pairing technique used
for estimating anammox and denitrification activities ............................................ 7

FIGURE 1.3. Typical porewater nitrogen profile from a deep ocean
sediment indicating a possible zone of anaerobic ammonium oxidation ...................... 9

FIGURE 1.4. Illustration of increasing relative importance of anammox
((anammox/(anammox+denitrification)) with increasing depth in sediments
up to 700m depth below water surface. Adapted from Dalsgaard et al., 2005. ..............10

FIGURE 2.1. 16S rRNA gene-based bootstrapped phylogenetic tree reflecting the
relationship of Brod541F-Brod126OR clones from East Hanna Shoal (EHS),

East Hanna Shoal Deep (Deep), Washington margin (Wash), Turning Basin (Turn),
Shallow Bud Inlet (Bud), Juan de F uca (F uca), Wintergreen Lake (Win), Sherrifi‘s
Marsh (She), Everglades F 1 (F1), Everglades U3 (U3), and Siberian permafrost

(Perm) sediments with 98% redundancy removed with the anammox group

and related bacteria given as GenBank accession numbers. The bar indicates

a 5% sequence divergence .......................................................................... 28

FIGURE 2.2. Neighbor-joining tree based on the 541F/1260R primer set

derived 16S rRNA gene sequences afier identical sequences were removed,

illustrating the lack of phylogenetic distinctiveness between freshwater and

marine samples. Right to left, top to bottom, with double bars indicating cutoff

to the next tree. Groupings of Michigan sediments and Everglades samples are

noted by triangles. DEEP=East Hanna Shoal Deep, Balt=four clone sequences

derived from amplification of Baltimore Harbor sediments by Tal et al., 2005,
Contract0F16S rRNA sequences derived from a rotating biological

contractor ............................................................................................. 29

FIGURE 2.3 Amplicons resulting from the An7F/1388R primer set in Shallow

Bud Inlet, East Hanna Shoal, and Washington Margin samples. lkb ladders were
used ..................................................................................................... 34

xi

FIGURE 2.4. 16S rRNA gene-based bootstrapped neighbor joining tree reflecting

the relationship of full length An7F-An1 388R generated amplicons with the

anammox group as well as other sequences derived from environmental samples

given as GenBank accession numbers. The bar indicates a 5% sequence

divergence ............................................................... . .............................. 36

FIGURE 3.1. Map of the eight station locations (S-S8) in the Cascadia Basin,
North East Pacific Ocean ................................................................................................... 47

FIGURE 3.2. Figure 3.2. Figure illustrating the experimental setup for the
determination of anammox activity ............................................................... 49

FIGURE 3. 3a. Pore water distribution of NH4 (A), N02 (0) and N03 (O) at
stations S1 -S4. At S2 and S4, Mn2+ (A) profiles are also shown. Note that
the scale for ammonium and Mn2+ concentrations vary between stations ......................... 54

'FIGURE 3. 3b. Pore water distribution of NH4 (A), N02 (.) and N03 (0) at
stations S5-S8. At S6 and S8, Mn:+ (A) profiles are also shown. Note that
the scale for ammonium and W2 concentrations vary between stations ....................... 55

FIGURE 3. 4. Production of 29N2 and 30N2 over time in anoxic sediment

incubations following additions of 40 nmol 15NH4 ml’l (wet sediment),

80 nmol 15NH4 + 40 nmol 14NO2 ml (wet sediment) and 40 nmol 15NO2

ml (wet sediment), n=2. Data are from 5N incubations at station S5,

anammox and denitrification rates were estimated from 29N2 and 30N2

production rates in the 15N02 treatment ........................................................................... 57

FIGURE 3.5a. Absolute anammox (white bar) and denitrification (black bar)

rates estimated from anoxic 15N sediment incubations. Error bars denote standard
deviation at the stations where a replicate core could be tested (S4, S5, and S8).

Percent values show anammox contribution to total N2 production at each depth

where N2 production rates were measured. Pore water nitrate distribution (0)

indicate the suboxic zone at each site. In this study the suboxic zone is defined

as the zone below oxygen penetration depth where nitrate is present ............................... 58

FIGURE 3.5b. Absolute anammox (white bar) and denitrification (black bar) rates
estimated from anoxic lsN sediment incubations at stations S5-S8. Error bars denote
standard deviation at the stations where a replicate core could be tested (S5 and S8).
Percent values Show anammox contribution to total N2 production at each depth

where N2 production rates were measured. Pore water nitrate distribution (0)

indicate the suboxic zone at each site. In this study the suboxic zone is defined

as the zone below oxygen penetration depth where nitrate is present .............................. 59

xii

FIGURE 3.6. Absolute denitrification (o) and anammox (O) rates in Cascadia

Basin distributed with distance from shore (km). Rates presented in the

diagrams are estimated from the top 2 cm of the suboxic zone (the zone below

oxygen penetration depth where nitrate is present). Distances were estimated

with the GPS program Tsunami 99 from Transas ............................................................. 60

FIGURE 3.7. Variable region 4 16S rRNA based UPGMA tree using sequences
obtained from 454 pyrosequencing in station 2 (0-2 cm depth) sediment .................... 63

FIGURE 3.8. Output of SDS software for a standard curve using a clonal full
length Scalindua brodae 16S rRNA gene with replication .................................... 64

FIGURE 3.9. Calibration curve linear regression of anammox 16S rRNA
gene copy numbers determined from a clone dilution series from 8 to 8x108
copies. CT is the threshold cycle number ......................................................... 65

FIGURE 3.10. Percent recovery of 10"4 added 16S rRNA gene copies
from each sediment matrix .......................................................................... 66

FIGURE 3.11. Q-PCR determined anammox copy number pg" environmental
DNA, g'1 wet sediment, and total estimated genomes g’l wet sediment at
each station ............................................................................................ 67

FIGURE 3.12. Linear regression of Q-PCR determined anammox gene copy ,ug'1
environmental DNA versus anammox N2 production. Data reflects average copy
numbers per 2 cm slice used for the determination of N2 production ........................ 68

FIGURE 3.13. Depth relationship between anammox copy number, NO2',
NHX, and N03' at station 2 .......................................................................... 69

FIGURE 4.1. Sampling site locations: l-East Hanna Shoal, 2-Barrow Canyon,
3-Juan de Fuca, 4-Washington Margin, 5-Cascadia Basin, 6-two Gulf of
Mexico samples, 7-Florida Bay 11, 8-Florida Bay 10, 9-Florida Bay 9 ...................... 86

FIGURE 4.2. Comparison of relative abundances based on the RDP Classifier
between the two most contrasting estimated richness and diversity sites,
Florida Bay 11 and Barrow Canyon ............................................................... 91

FIGURE 4.3. Phylogenetic architecture based on a normalized sample size of
3,500 randomly selected sequences per site. Dissimilarity values for species,
genus, and family are noted ........................................................................ 92

FIGURE 4.4. Relative abundances illustrating the phylogenetic composition

of the sequences obtained at each site based on the RDP Classifier at
50% bootstrap confidence threshold ............................................................... 95

xiii

FIGURE 4.5. NEO plot of RDP best hit matches from all samples to isolates

with S_ab score by phylum. One column represents a best match to one isolate.
Percentage based frequency distribution highlighted on the right. Dashed line

indicates an estimated 85% pairwise identity based on the S_ab score ...................... 96

FIGURE 4.6. Frequency distribution of S_ab scores illustrating differences
in isolate coverage among three sites. Dashed line indicates an estimated
85% pairwise identity based on S_ab scores ..................................................... 97

FIGURE 4.7. BLAST best match scores for each cluster representative sequence
based on the cluster representative sequence RDP classifier results .......................... 98

FIGURE 4.8. Chao-Sorensen cluster abundance based UPGMA tree at 5%
dissimilarity rooted with an agricultural soil from Ghana with (A) all
clusters and (B) without singletons. Depth below water surface is noted .................... 99

FIGURE 4.9. Relationship of identified sequences with >98% sequence
homology to the known anammox bacteria. Site designations are referred
at the beginning of each sequence name ........................................................ 100

FIGURE 5.1. Standard curve based on a dilution series of plasmids
containing the 16S rRN A gene of Brocadz'a anammoxidans. ............................... 116

FIGURE 5.2. Distribution of all anammox bacteria based on the 818F/1066R

primer set (dark grey) compared to the marine only Scalindua sp. related

anammox (light grey). Abundances are normalized to 16S rRNA gene copy

number g'1 wet sediment'l ......................................................................... 117

xiv

CHAPTER I
GENERAL INTRODUCTION

The anammox reaction is anaerobic oxidation of ammonium coupled with nitrite
reduction under anoxic conditions. This alternative nitrogen removal pathway was first
proposed by Richards (1965), following observations of that indicated an ammonium
deficit in anoxic marine basins. Throughout most of the twentieth century ammonium
was believed to be inert under anoxic conditions. Canonical denitrification liberates
ammonium from organic matter during respiration, resulting in net accumulation in the
sediment/soil profile. The proposed ‘anammox’ pathway allows for the removal of
ammonium under purely anoxic conditions. Early evidence for the presence of this
reaction was provided by marine sediment porewater profiles where the simultaneous
disappearance of nitrite and ammonium was observed (Codispoti and Richards 1976;
Cline and Richards 1972). Broda (1977) soon proposed a new type of bacteria
responsible for these observations, a “chemosynthetic bacteria that oxidize ammonia to
nitrogen with 02 or nitrate as an oxidant”, which was coined one of two “lithotrophs
missing in nature”. It was not until 1995 that the anammox process was confirmed in a
fluidized bed reactor treating wastewater effluent (Mulder et a1. 1995). The anammox

reaction is a chemolithotrophic process in which 1 mole of ammonium is oxidized by 1

mole of nitrite to produce N2 gas in the absence of oxygen (Strous et a1. 1999):
NH4+ + NO2' —» N2 + 21120

Compared to denitrification, this process produces twice the amount of N2 per

mole nitrite consumed and increases N2 production in sediments where nitrification is

limited. The bacteria responsible for this process were later identified as a deep-
branching planctomycete with a peculiar morphology (Strous et al. 1999). Despite the
early connotation that many microbes cannot be isolated in pure culture (Winogradsky
1949), the presence of a microbially-mediated reaction that disputed the notion that
ammonium was inert under anoxic conditions was initially regarded with skepticism.
Since then, numerous studies have identified anammox as a key process in the global

nitrogen cycle.

Anammox physiology and metabolism

All currently known bacteria capable of anaerobic ammonium oxidization belong to
a deep branching lineage of the order Planctomycetales with high genus level diversity
(Freitag and Prosser 2003; Schmid et al. 2003). The evolutionary distance among the
anammox genera are large (<85% 16S rRNA gene nucleotide identity) though they share
the same basic anammox metabolism and cell structure. There are currently four
Candidatus genera whose grouping are largely based on 16S rRNA sequences: The
“freshwater” Kuenenia (K. stuttgartiensis; Schmid et al. 2000) and Brocadia (B.
anammoxidans (5) and B. fulgida (22)), and the “marine” anammox Scalindua (S.
‘ sorokinii, S. brodae, and S. wagneri; Schmid et al. 2003). The fourth Candidatus genus
has one member, Anammoxoglobus propionicus (Kartal et al. 2007b) which exhibits an
alternative metabolism. Anammox bacteria are characterized by a membrane-bound
organelle called the anammoxosome that comprises more than 30% of the cell volume.
This intracytoplasmic compartment is surrounded by unique lipids, called ladderanes

(Sinninghe Damsté et al. 2002) that are unique to the anammox bacteria. Ether and ester

linkages tie the lipids to a glycerol backbone in the membrane which have historically
only been found in members of the domain Archaea and may reflect an early divergence
of anammox in the bacterial lineage (Brochier and Philippe 2002). Due to a very dense
arrangement of carbon atoms, the ladderane lipids serve as a diffusion barrier (Sinninghe
Damsté et al. 2002). This may serve to protect the bacteria from the toxic anammox
reaction intermediates hydroxylamine and hydrazine (Jetten et al. 2003). Due to their
unique characteristics, ladderane lipids have also been used as a biomarker for the
presence of anammox bacteria (Kuypers et al. 2003).

Evidence from the genome of Candidatus K. stuttgartiensis (Strous et al. 2006)

indicates that the anammox reaction proceeds via the following steps:
N02' ——> NO
NO + NH4+ —'> N2H4 ->N2

The anammox hydroxylarnine oxidoreductase (HAO) enzyme is responsible for the
oxidation of hydrazine to N2 gas and is located exclusively within the anammoxosome
(Lindsay et al. 2001), a possible target for future molecular studies. The highly reactive
hydrazine intermediate is stored inside the anammoxosome (Sinninghe Damsté et al.
2002), which is especially important considering the slow enzymatic turnover, resulting
in a doubling time of nine days in optimal conditions for the “freshwater” anammox
(Strous et al. 1999). Anammox are reversibly inhibited by 02 and reaction rates are the
same after as before aeration (J etten et al. 1999).

Anammox bacteria have been found to be metabolically flexible, exhibiting
alternative metabolic pathways. For instance, anammox can reduce nitrate to nitrite to

ammonium, followed by the conversion of ammonium and nitrite to N2 through the

anammox pathway, allowing anammox bacteria to overcome ammonium limitation.
Anammox bacteria are also a potential minor source N20 production by nitric oxide
detoxification (Kartal et al. 2007a). Currently the other known processes that produce
N20 are nitrification and denitrification (Fig. 1.1). As such, classical denitrification
measures that depend exclusively on N20 measures may overstate the role of
denitrification in the system. Specifically, the acetylene block technique does not detect
anammox N2 production, as N20 is not a major intermediate of the reaction. Another
alternative pathway is carried out by Candidatus Anammoxoglobus propionicus, which
has been shown to co-oxidize propionate and ammonium and out-compete denitrifiers
and other anammox bacteria in the process (Kartal et al. 2007b). This supports the niche
differentiation of anammox in which different “ecotypes”, genetically unique populations
adapted to their local environment. Lastly, iron and manganese oxides have also been
found to be respired with forrnate as an electron donor (Strous et al. 2006), further
expanding the metabolic diversity of the anammox bacteria.

This metabolic flexibility is also evident within the Candidatus Scalindua marine
anammox which were able to use both formate and acetate as electron donors and
manganese and iron as electron acceptors (van de Vossenberg et al., 2008). Temperature
also appears to be a selective condition between two different closely related Scalindua
phylotypes in enrichment cultures. This supports the findings that specific anammox 16S

rRNA based phylotypes generally dominate in a given sediment or regional area.

Detection of anammox bacteria and activity

The isotope pairing technique (IPT) has been used as the standard measure of anammox

activity, most commonly using homogenized sediments (Thamdrup and Dalsgaard 2002).

Aerobic

‘3‘ N02- Organic N

 

 

 

 

Canonica1 Denimficatron Anaerobic

Figure 1.1. Anaerobic ammonium oxidation pathway of N removal in context of the
nitrogen cycle.

Concentrations of NH4+, N03", and N02' are first determined, the sediments are placed in
airtight containers with a septum, such as Exetainer tubes, and the headspace is flushed
with He for a minimum of five minutes to replace ambient 02. Concentrations of residual
N03' and N02' are monitored over time until all available oxidized species is removed
from the incubations. Three parallel incubations are then performed: (1) 15 NH] alone, (2)

the combination of 15NH4+ and 14N02‘, and (3) 15N02' alone.

Reactions are stopped by the addition of ZnCl2.The first incubation is used as a
control to detect any oxidation of ammonium without the addition of nitrite. The lack of
29N2/30N2 is indicative of the lack of oxidants at the end of the pre-incubations. The
second treatment is used to determine if anammox activity is possible. The production of
29N2 indicates anammox activity through the oxidation of ammonium with nitrite. The
combination of the first two incubations is used to establish anammox activity. Finally,
the third incubation is used to estimate anammox and denitrification rates (Figure 1.2).
Anammox produces 2C)N2 through the oxidation of the resident NI-I4+ pool with the added
15N02' while denitrification is measured by the production of 30N2. However, evidence
that anammox can also reduce 15N03' to 15N02' to ”NHX (Kartal et al. 2007a) results in
the possibility that the anammox reaction can pair 15’N02' with ”NHL and thus some
proportion of measured denitrification may be partitioned to the anammox reaction.
Several modifications to this protocol are promising, notably the addition of N20
measures to more accurately quantify N2 production and the use of intact sediment cores
(Trimmer et al. 2006).

Molecular methods have been extensively utilized to identify the presence of
anammox bacteria in environmental and wastewater samples. Fluorescence in-situ
hybridization (FISH) targeting the 16S rRNA gene has been used extensively and is
described in detail by Schmid et al. (2005). Anammox bacteria have also been identified
using PCR with a variety of primers, often based on FISH probes, targeting the group as a
whole or specific members (Schmid et a1. 2005; Penton and Tiedje 2006). The unique
ladderane lipids that constitute the anammoxosome have also been used as biomarkers for

relative quantification (Kuypers et al. 2003) while distinctive hopanoid lipids may be

   

Isotope Ratio GC-MS

_ 29
29/30N2 = 29N2 = Anammox N2
AmbleIIt NO or Anammox Denitrification = 30N

N02 present P0551ble 2

Figure 1.2. Experimental layout of the isotope pairing technique used for estimating
anammox and denitrification activities.

useful in assessing relative anammox abundance in the sedimentary record (Sinninghe
Damsté et al. 2004). Quantitative PCR (q-PCR) has been used for direct quantification of
all known anammox-like bacteria in water columns (Hamersley et al. 2007) and

wastewater enrichment cultures (Tsushima et al. 2007).

Anammox in the environment

The linkage of anammox activity with the removal of fixed inorganic nitrogen in
natural systems was first confirmed in the Black Sea suboxic water column (Kuypers et
al. 2003). Since then anammox has been shown to be a significant contributor to nitrogen
losses in a variety of environments, responsible for 19-35% of the nitrogen loss in an
anoxic coastal bay (Dalsgaard et a1. 2003) and the majority of N removal in one of the

most productive regions of the world’s oceans, the Benguela upwelling oxygen minimal

zone (Kuypers et al. 2005). These sites exhibit characteristics of oxygen minimum zones,
which are thought to be responsible for 30 to 50% of global N removal (Brandes and
Devol 2002). Evidence for the anammox reaction in sediments or soils is generally first
determined by the porewater N profile. Anoxic zones where there is a concomitant
reduction in both nitrite/nitrate and ammonium exhibits the initial conditions necessary
for anammox activity (Figure 1.3). The maximum reported contribution of anammox is
67-79%, occurring in sediments at a depth of 700m (Engstrdm et al. 2005) which led to
the hypothesis that relative anammox contributions increase with depth. Overall trends
suggest that while anammox N2 production decreases with depth, the total importance of
this process in relation to denitrification increases with depth (Figure 1.4). However,
these data only cover sediments ranging from 0m to 700m and do not address deeper
sediments where both organic matter recalcitrance and N02' / N03' availability may
affect the relationship between denitrification and anammox activity.

Ammonium is typically abundant in anoxic systems, provided by organic matter
oxidation. Nitrate reducing or aerobic ammonium oxidizing bacteria provide the nitrite
necessary for the anammox reaction. As such, organic matter availability is thought to be
a major factor influencing the relative significance of anammox to total N2 production.
Greater organic matter availability creates a higher demand by denitrifiers for N02' and
N03‘, and less N02' is liberated for anammox consumption. Nitrite is also an intermediate
product of denitrification and the end product of nitrate respiration. The competition for
N02' between denitrifiers and anammox bacteria shows a non-linear response. In
particular, at low N02' concentrations (2-5 uM), anammox bacteria take proportionally

more of the available nitrite due to a high N02' affinity (Trimmer et al., 2005). Therefore,

 

 

 

 

 

 

Figure 1.3. Typical porewater nitrogen profile from a deep ocean sediment
indicating a possible zone of anaerobic ammonium oxidation. Figure based on data
from Chapter III.

A/(A+D)%
o 20 4o 60 80

Om

0.5 m

10-20 m

20-30 m

30-40 m

40-50 m

50-100 m

100-500 In
500-700 m

 

Figure 1.4. Illustration of increasing relative importance of anammox
((anammox/(anammox+denitrification)) with increasing depth in sediments up to
700m depth below water surface. Adapted from Dalsgaard et al., 2005.

10

in deep marine sediments, where low N02' concentrations dominate due to decreased
organic matter availability, anammox would be expected to dominate N02' removal.

Anammox contributed O to 9% of total N2 production in subtropical mangrove
sediments (Meyer et al. 2005), 8% in estuarine sediments (Trimmer et al. 2003), and less
than 2% in eutrophic shallow coastal bay sediments (Thamdrup and Dalsgaard 2002).
Although sediment reactivity has been negatively correlated with anammox contribution
to total N2 production (Trimmer et al. 2003), absolute anammox rates appear to peak at
sites with intermediate reactivity (Engstrdm et al. 2005). Thus, a strict relationship
between anammox activity and organic matter availability is not firmly established.
Additionally, due to the slow growth of anammox and their inhibition by low
concentrations of 02 (if the “marine” anammox respond the same as the “freshwater”
species), environmental stability may be an important controlling factor of anammox
activity.

16S rRNA sequences identical or closely related to the marine anammox have
been found widely distributed in marine systems, freshwater lakes, and subtropical
wetlands (Penton and Tiedje 2006). Anammox 16S rRNA sequences have also been
identified in Siberian frozen alluvial sandy loam, deposited in the Middle Pleistocene
Epoch 300,000-400,000 years ago in the Cape Svyatoi Nos tundra zone on the Laptev
Sea coast (Penton and Tiedje 2006). Anammox activity was responsible for up to 19% of
total N2 production in a stable Greenland Sea ice floe but was not detectable in annual sea
ice (Rysgaard and Glud 2004). Both aerobic and anaerobic processes in microzones were
found to occur simultaneously in brine pockets. This raises the possibility that anammox

contributes to N2 removal in permafrost soils.

11

Relatively few studies have investigated anammox activity in natural freshwater
systems, although the “freshwater” anammox bacteria are the most intensively
characterized due to their implementation in wastewater treatment bioreactors. Schubert
et al. (2006) reported an anammox contribution of 13% in the largest freshwater anoxic
lake in the world, Lake Tanganyika. Anammox 16S rRNA gene sequences with <96%
sequence identity to Candidatus Scalindua brodae were identified in the anoxic water
column and anammox cells were enumerated using FISH. Molecular analysis was used
to assess the diversity of the anammox population in the Xinyi River (China) (Zhang et
al. 2007). Sequences, obtained by targeted PCR, exhibited 16S nucleotide identities of
95% to Candidatus Brocadia anammoxidans and 95% to the Candidatus Scalindua
species, including the sequence obtained from the Lake Tanganyika study. These findings
suggest that more diverse anammox communities may exist in freshwater habitats,

compared to the multitude of marine studies that indicate a single, dominant anammox

phylotype.

OBJECTIVES

I undertook this work to address the following objectives and hypotheses:
Objective 1

To study the distribution and diversity of anammox bacteria in a broad

range of biogeochemical conditions and geographic locations.

At the point that this study was undertaken anammox bacteria had only been

identified in wastewater treatment plants (wwtps) and anoxic water columns. The

12

distribution of the “marine” Scalindua sp. type anammox was largely unknown, and was
thought to be constrained to obligately anoxic environments, sufficiently stable to harbor
a slow growing consortia. Therefore, it was necessary to establish what types of
environments in which these bacteria may be identified as a first step in understanding
their potential importance on a global scale. Primers used for anammox identification
were generally based on the Planctomycetes specific primers, which exhibited a low
specificity for anammox bacteria. It was necessary to develop new primers that could be
used as a screening tool for the specific identification of anammox bacteria. Although
presence does not necessarily indicate significant functionality, a widespread anammox
distribution would provide evidence that anaerobic ammonium oxidation was not
confined to a very selective environment. Based on these data, I tested the following

hypothesis:

Hypothesis 1

Anammox bacteria will be identified in many conditions that would not appear to
harbor a stable, anoxic environment due to anoxic microsites within sediment particles
that would be supportive of an anammox metabolism while the sample, in general, is not

a stable anoxic environment.

13

Objective 2
To develop a quantitative PCR method for the rapid enumeration of
anammox bacteria in marine sediments in order to determine the relationship

between abundance and activity.

The enumeration of anammox bacteria is based on the use of fluorescence in-situ
hybridization (FISH), which was described as the “gold standard”. However, the
implementation of FISH in sediments is especially difficult due to obstacles such as non-
specific hybridization and particle shielding effects, among others. One study that
attempted to use FISH in sediments appeared to drastically undercount these bacteria
when cell-specific activity was calculated (Tal et al., 2005). Therefore, it was necessary
to develop a quantitative PCR method that would specifically target the 16S rRNA gene
of the marine Scalindua sp. type anammox. Estimated abundances will then be directly
related to anammox activity obtained through 15N assays. The following exclusive

hypotheses were tested:

Hypothesis 1:

Anammox bacteria will only be identified in these stable, anoxic environments
where oxidized nitrogen species are identified.

Hypothesis 2:

Alternatively, anammox will be identified in depths where oxidized nitrogen

species are not present. This may be due to bioturbation creating preferential porewater

14

channels, transient pulses of NO2' / N03' that may penetrate deeper in the sediment

column, or may reflect alternative metabolic pathways of anammox bacteria.

Objective 3
To determine the effect of increasing depth below water surface on anammox
activity in order to test the hypothesis that anammox is the prevalent N2 removal

pathway in deep marine sediments.

The relative importance of anammox, compared to denitrification, showed an
overall increase with increasing sediment depth below water surface (Thamdrup and
Dalsgaard 2002; Risgaard-Petersen et al. 2004; EngstrOm et al. 2005). AS such, it was
thought that anammox would be responsible for the majority of nitrogen removal in
depths greater than 1000m (Dalsgaard et al., 2005). This was thought to be due to
decreasing reactive organic matter availability, which would favor anammox over
denitrification (Thamdrup and Dalsgaard 2002; Engstrbm et al. 2005). The following

hypothesis were tested:

Hypothesis 1:
Anammox activity will not be a dominate pathway in N removal in deep marine
sediments. The very low organic matter availability of deep sea sediments will limit the

supply of N02' / N03' to anammox bacteria.

15

Objective 4
To investigate differences in microbial community structure and diversity
among marine sediments obtained in a variety of biogeochemical and geographic

locations.

Surveys of microbial communities in a wide range of habitats have primarily
depended on clone library analysis of 16S rRNA gene libraries. The lack of sufficient
resolution or depth in many of these studies did not allow for adequate comparisons to be
made, especially in regard to the rarer community members. Understanding community
diversity among widely ranging geographical locations is important, especially in relation
to the dogma that “everything is everywhere, the environment selects”. The controls on
diversity are not understood but may include environmental stability, biogeochemistry,
temperature, salinity, or type of carbon input. In this study the following hypotheses were

tested:

Hypothesis 1
Marine sediments from deep sites will exhibit a more similar community
structure, due to the higher processing of carbon as it sinks from the pelagic zone, which

reflects a more temporally stable environment.
Hypothesis 2
Community structures will be more similar in sites that share a common marine

basin (e.g. Pacific sites vs. Gulf of Mexico sites).

16

REFERENCES

Brandes J.A. and A.H. Devol. 2002. A global marine-fixed nitrogen isotopic budget:
Implications for Holocene nitrogen cycling. Global Biogeochem. Cycles 16:1120-1133.

Brochier C. and H. Philippe. 2002. Phylogeny—a non hypertherrnophilic ancestor for
bacteria. Nature. 417:244.

Broda E. 1977. Two kinds of lithotrophs missing in nature. Z. Allg. Mikrobiol. 17:491-
493.

Cline J.D. and EA. Richards. 1972. Low oxygen concentrations and nitrate reduction in
the eastern tropical North Pacific Ocean. Limnol. Oceanogr. 17:885-900.

Codispoti L.A. and EA. Richards. 1976. An analysis of the horizontal regime of
denitrification in the eastern tropical North Pacific. Limnol. Oceanogr. 21 :379-388.

Dalsgaard T., Canfield D.E., Petersen J., Thamdrup B., and J. Acur‘ra-Gonzalez.
2003. N2 production by the anammox reaction in the anoxic water column of Golfo
Dulce, Costa Rica. Nature. 422:606-608.

Dalsgaard T., Thamdrup B., and D.E. Canfield. 2005. Anaerobic ammonium
oxidation (anammox) in the marine environment. Res. Microbiol. 156:457-464.

Engstrtim P., Dalsgaard T., Hulth S., and RC. Aller. 2005. Anaerobic ammonium
oxidation by nitrite (anammox): Implications for N2 production in coastal marine
sediments. Geochim. Cosmochim. Acta. 69:2057-2065.

Freitag TE. and JJ. Prosser. 2003. Community structure of ammonia oxidizing
bacteria within anoxic marine sediments. Appl. Environ. Microbiol. 69: 1359—1371.

Hamersley M.R., Lavik G., Woebken D., Rattray J.E., Lam P., Hopmans E.C.,
Sinninghe Damsté J.S., Kriiger S., Graco M., Gutierrez D., and M.M.M. Kuypers.
2007. Anaerobic ammonium oxidation in the Peruvian oxygen minimum zone. Limnol.

Oceanogr. 52:923-933.

Jetten M.S.M., Sliekers 0., Kuypers M., Dalgaard T., van Niftrik L., Cirpus 1., van
de Pas-Schoonen K., Lavik G., Thamdrup B., Le Paslier D., Op den Camp H.J.M.,
Hulth S., Nielsen L.P., Abma W., Third K., Engstriim P., Kuenen J.G., Jorgensen
B.B., Canfield D.E., Sinninghe Damsté J.S., Revsbech N.P., F uerst J., Weissenbach
J., Wagner M., Schmidt 1., Schmid M., and M. Strous. 2003. Anaerobic ammonium
oxidation by marine and freshwater planctomycete-like bacteria. Appl. Microbiol.
Biotechnol. 63: 107-1 14.

17

_ Jetten M.S.M., Strous M., van de Pas-Schoonen K.T., Schalk J., van Dongen
U.G.J.M., van de Graaf A.A., Logemann S., Muyzer G., van Loosdrecht M.C.M.,

and J.G. Kuenen. 1999. The anaerobic oxidation of ammonium. FEMS Microbiol Rev
22:421-437.

Kri‘rger S., Graco M., Gutierrez D., and M.M.M. Kuypers. 2007. Anaerobic
ammonium oxidation in the Peruvian oxygen minimum zone. Limnol Oceanogr 52:923-
933.

Kartal B., Kuypers M.M.M., Lavik G., Schalk J., Op den Camp H.J.M., Jetten
M.S.M., and M. Strous. 2007a. Anammox bacteria disguised as denitrifiers: nitrate
reduction to dinitrogen gas via nitrite and ammonium. Environ Microbiol 9:635-642.

Kartal B., Rattray J., van Niftrik L.A., van de Vossenberg J., Schmid M.C., Webb
R.I., Schouten S., Fuerst J.A., Sinninghe Damsté J., Jetten M.S.M., and M. Strous.
2007. Candidatus “Anammoxoglobus propionicus” a new propionate oxidizing species of
anaerobic ammonium oxidizing bacteria. Syst Appl Microbiol 30:39-49.

Kuypers M.M.M., Slickers A.O., Lavik G., Schmid M., Jorgensen B.B.,Kuenen J.G.,
Sinninghe Damsté J.S., Strous M., and M.S.M Jetten. 2003. Anaerobic ammonium
oxidation by Anammox bacteria in the Black Sea. Nature 422:608—61 1.

Kuypers M.M.M., Lavik G., Woebken D., Schmid M., Fuchs B.M., Amann R.,
Jorgensen B.B., and M.S.M. Jetten. 2005. Massive nitrogen loss fi'om the Benguela

upwelling system through anaerobic ammonium oxidation. Proc Natl Acad Sci 10226478-
6483.

Lindsay M.R., Web R.I., Strous M., Jetten M., Butler M.K., and J.A. Fuerst. 2001.
Cell compartmentalization in planctomycetes: novel types of structural organization for
the bacterial cell. Arch Microbiol 175:413—429.

Meyer R.L., Risgaard-Petersen N., and D.E. Allen. 2005. Correlation between
anammox activity and microscale distribution of nitrite in a subtropical mangrove
sediment. Appl Environ Microbiol 71 :6142-6149.

Mulder A., Vandegraaf A.A., Robertson L.A., and J.G. Kuenen. 1995. Anaerobic
ammonium oxidation discovered in a denitrifying fluidized-bed reactor. FEMS Microbiol
Ecol 16:177—183.

Richards EA. 1965. Anoxic basins and fjords. In: Riley JP, Skirrow G (eds) Chemical
Oceanography, vol 1. Academic Press, London, pp 611-645.

Rysgaard S. and RN. Glud. 2004. Anaerobic N2 production in Arctic sea ice. Limnol
Oceanogr 49:86-94.

18

Schmid M., Twachtmann U., Klein M., Strous M., Juretschko S., Jetten M.S.M.,
Metzgner J., Schleifer K.H. and M. Wagner. 2000. Molecular evidence for genus level
diversity of bacteria capable of anaerobic ammonium oxidation. Syst Appl Microbiol
23:93-96.

Schmid M., Walsh K., Webb R., Rijpstra W.I.C., van de Pas- Schoonen K.T.,
Verbruggen M.J., Hill T., Moffett B., Fuerst J., Schouten S., Sinninghe Damste' J.S.,
Harris J., Shaw P., Jetten M.S.M., and M. Strous. 2003. Candidatus “Scalindua
brodae,” sp. nov., Candidatus “Scalindua wagneri”, sp. nov., two new species of
anaerobic ammonium oxidizing bacteria. Syst Appl Microbiol 26:529—538.

Schmid M.C., Maas B., Dapena A., van de Pas-Schoonen K., van de Vossenberg J.,
Kartal B., van Niftrik L., Schmidt 1., Cirpus 1., Kuenen J.G., Wagner M., Sinninghe
Damsté J.S., Kuypers M., Revsbech P., Mendez R., Jetten M.S.M., and M. Strous.
2005. Biomarkers for the in situ detection of anaerobic ammonium-oxidizing (anammox)
bacteria. Appl Environ Microbiol 71 :1677-1684.

Schubert C.J., Durlsch-Kalser E., Wehril B., Thamdrup B., Lam P., and M.M.M.
Kuypers. 2006. Anaerobic ammonium oxidation in a tropical freshwater system (Lake
Tanganyika). Environ Microbiol 8: 1857-1863.

Sinninghe Damsté J.S., Strous M., Rijpstra W.I.C., Hopmans E.C., Geenevasen
J.A.J., Van Duin A.C.T., Van Niftrik L.A., and M.S.M. Jetten MSM. 2002. Linearly
concatenated cyclobutane lipids form a dense bacterial membrane. Nature 419:708—712.

Sinninghe Damsté J.S., Rijpstra W.I.C., Schouten S., Fuerst J.A., Jetten M.S.M.,
and M. Strous. 2004. The occurrence of hopanoids in planctomycetes: implications for
the sedimentary biomarker record. Org Geochem 35:561—566.

Strous M., F uerst J.A., Kramer E.H.M., Logemann S., Muyzer G., van de Pas-
Schoonen K.T., Webb R., Kuenen J.G., and M.S.M. Jetten. 1999. Missing lithotroph
identified as new planctomycete. Nature 400:446—449.

Strous M., Kuenen J.G., Jetten M.S.M. 1999. Key physiology of anaerobic ammonium
oxidation Appl Environ Microbiol 65:3248—3250.

Strous M., Pelletier E., Mangendt S., Rattei T., Lehner A., Taylo M.W., Horn M.,
Daims H., Bartol-Mavel D., Wincker P., Barbe V., Fonknechten N., Vallenet D.,
Segurens B., Schenowitz-Truong C., Médigue C., Collingro A., Snel B., Dutilh B.E.,
Op den Camp H.J.M., van der Drift C., Cirpus I., van de Pas-Schoonen K.T.,
Harhangi H.R., van Niftrik L., Schmid M., Keltjens J., van de Vossenberg J., Kartal
B., Meier H., Frishman D., Huynen M.A., Mewes H.-W., Weissenbach J., Jetten
M.S.M., Wagner M., and D. Le Paslier. 2006. Deciphering the evolution and
metabolism of an anammox bacterium from a community genome. Nature 440:790—794.

19

Thamdrup B., and T. Dalsgaard. 2002. Production of N2 through anaerobic ammonium
oxidation coupled to nitrate reduction in marine sediments. Appl Environ Microbiol
68:1312-1318.

Trimmer M., Risgaard-Petersen N., Nicholls J.C., and P. Engstrom. 2006. Direct
measurement of anaerobic ammonium oxidation (anammox) and denitrification in intact
sediment cores. Mar Ecol Prog Ser 326:3 7-47.

Trimmer M., Nicholls J.C., Deflandre B. 2003. Anaerobic ammonium oxidation

measured in sediments along the Thames estuary, United Kingdom. Appl Environ
Microbiol 69:6447—6454.

Tsushima 1., Kindaichi T., Okabe S. 2007. Quantification of anaerobic ammonium-
oxidizing bacteria in enrichment cultures by real-time PCR. Water Res 41 :785-794.

Winogradsky S. 1949. Principes de la Microbiology oecologique. Microbiologie du Sol.
Masson et cie, Paris.

Zhang Y., Ruan X.-H., Op den Camp H.J.M., Smits T.J.M., Jetten M.S.M., and
M.C. Schmid. 2007. Diversity and abundance of aerobic and anaerobic ammonium-

oxidizing bacteria in freshwater sediments of the Xinyi River (China). Environ Microbiol
9:2375-2382.

20

CHAPTER II

MOLECULAR EVIDENCE FOR THE BROAD DISTRIBUTION OF ANAMMOX

IN FRESHWATER AND MARINE SEDIMENTS

ABSTRACT

Previously available primer sets for detecting anammox bacteria are inefficient,
resulting in a very limited database of such sequences which limits knowledge of their
ecology. To overcome this, I designed a new primer set that was 100% efficient in
recovering ~700 bp 16S rRNA gene sequences with >96% homology to the "Candidatus
Scalindua" group of anammox, and detected this group in all sites studied, including a
variety of freshwater and marine sediments and permafroét soil. A second primer set was
designed that exhibited greater efficiency than previous primers in recovering full length
(1380 bp) sequences related to "Ca. Scalindua," "Ca. Brocadia," and "Ca. Kuenenia."
This study provides evidence for the widespread distribution of anammox in that it
detected closely related anammox 16S rRNA gene sequences in eleven geographically

and biogeochemically diverse freshwater and marine sediments.

INTRODUCTION

Anaerobic ammonium oxidation (anammox) involves the direct autotrophic
conversion of ammonium and nitrite to N2 under anoxic conditions. First confirmed in a

fluidized wastewater bed reactor in 1994 (Mulder et al. 1995), anammox bacteria are now

21

thought to form a deep branching lineage of the Planctomycetales with a high genus level
diversity (Freitag et al. 2003, Schmid et al. 2003). The anammox reaction occurs within
a specialized cell compartment, the anammoxosome, in which ammonium is oxidized via
hydrazine (N2H4) and hydroxylamine (NH2OH) intermediates (Schalk et al. 1998). This
process produces twice the amount of N2 per molecule of nitrite consumed as
denitrification and does not require an external reducing agent (Kuai and Verstraete 1998,

Strous et al. 1999).

The role of anammox bacteria in the nitrogen cycle has been investigated in only
a few environments. Anammox has been proposed to be responsible for the consumption
of more than 40% of the fixed N in the anoxic water in the Black Sea (Kuypers et al.
2005); 19 to 35% of the total N2 formation in the anoxic waters of a coastal bay in Costa
Rica (Dalsgaard et al. 2003); between 24 and 67% of total N2 production in continental
shelf sediments (Dalsgaard et al. 2002) and between 1 and 24% of the N lost in estuarine
sediments (Risgaard-Petersen et al. 2004, Trimmer et al. 2003). In a global context, the
anammox reaction may account for 30 to 70% of total oceanic N2 production (Devol

2003), representing a major N sink.

While isotope pairing studies have been used as indicators of anammox
metabolism, the detection of anammox bacteria in environmental samples has been -
primarily by fluorescence in-situ hybridization (FISH). Few environmentally derived
anammox 16S rRNA gene sequences have been obtained by the use of "universal",
Planctomycetales targeted, or anammox-specific primers. Sequences have been
identified in relatively few, mostly marine, environments (Bowman and McCuaig 2003,

Kirkpatrick et al. 2006, Kuypers et al. 2003, 2005, Risgaard-Peteresen et al. 2004, Tal et

22

al. 2005) that fall within the documented anammox group. The previously available
primer sets for detecting anammox bacteria have resulted in a very limited database of
such sequences and may perhaps be limiting the knowledge concerning the distribution of

these important N cycling bacteria.

In this study I explored the distribution and diversity of anammox-related bacteria
in sediment samples from geographically and biogeochemically distinct environments.
Eleven samples from unique sites were analyzed, representing subtropical freshwater
wetlands, a northern fen wetland, a hypereutrophic lake, deep ocean, continental margin,

shallow marine, and Siberian permafrost.

MATERIALS AND METHODS

Sampling

Our sampling sites were chosen to represent contrasting geographical,
biogeochemical, and temporal attributes. Freshwater sediment samples were collected
from beneath Shallow (<1 m) waters at two long-term monitoring sites at Kellogg
Biological Station, Mich: Wintergreen Lake, a small, groundwater-fed, hypereutrophic
lake (42.399N 85.383W), and Sherriffs Marsh (42.422N 85.516W), a fen wetland that
receives groundwater and stream flow. Sediments from subtropical wetlands were
collected using a piston corer in the Florida Everglades Water Conservation Area 2A, at
two sites: a high phosphorus (P) eutrophic site F 1 (0.4 m depth, 26.366N 80.373W) and

a reference P level oligotrophic site U3 (0.5 m depth, 26.29ON 80.419W).

23

Marine samples were collected at six sites using a multicorer: East Hanna Shoal
(160 m depth, 72.637N 158.667W) and East Hanna Shoal Deep (1450 m depth, 72.852N
158.207W) both from the Alaskan maritime in the Chuckchi Sea; Turning Basin (12 m
depth, 47.088N 123.024W) and Shallow Bud Inlet (3 m depth, 47.341N 122.944W), both
in Puget Sound; Washington Margin (1138 m depth, 46.575N 124.822W) and West of
the Juan de Fuca Ridge (3869 m depth,46.783N 133.667W), both in the Pacific Ocean.
Siberian frozen alluvial sandy-loam deposited in the 300-400 thousand year ago Middle
Pleistocene Epoch was collected from Cape Svyatoi Nos on the Laptev Sea coast tundra

zone (72.917N 140.17E) as described previously (Vishnivetskaya et al. 2000).

Primer Design

Ten primers targeting the 16S rRNA genes of "Ca. Scalindua", "Ca. Brocadia",
and "Ca. Kuenenia" were initially designed, screened against GenBank for target
organisms, and tested in various sediments by PCR amplification and sequencing. The
following primer sets did not achieve consistent amplification: Wag991F (5’-
TAGCCAACGGTATCCAGTCC-3 ’) and Wagl 189R (5 ’ -
AAGGGCCATGATGACTTGAC—3’), targeting only “Ca. Scalindua wagneri”;
Brod54l F (5’-GAGCACGTAGGTGGGTTTGT-3’) and Stut766R (5’-
CGGGGTATCTAATCCCGTTT-3’), targeting Kuenenia stuttgartiensis; Sor0155F (5’-
TTGTCAATGGGAGGGTAGC-3 ’) and Sor03 3 1 R (5 ’ -

GATTCTCGACTGCAGCCTTC-3’), targeting “Ca. Scalindua sorokinii” and “Ca.

24

Scalindua brodae”. In addition, Brod541F, Wag99lF, and Sor0155F were combined with

the “universal” primer 1392R (5’-ACGGGCGGTGTGTAC-3’).

Brod541F (5'- GAGCACGTAGGTGGGTTTGT-3') and Brod126OR (5'-
GGATTCGCTTCACCTCTCGG-3') was the only primer set to consistently produce
single bands and specifically target anammox bacteria (720 bp amplicons). Specifically,
Brod541F exhibited 100% homology to four groups in GenBank; "Ca. Scalindua brodae"
(Schmid et al. 2003), uncultured ammonia oxidizing bacteria (Risgaard-Petersen et al.
2004), unidentified bacteria and planctomycetes from Antarctica (Bowman and McCuaig
2003), and uncultured planctomycetes from the Black Sea (Kirkpatrick et al. 2006).
Primers An7F (5'- GGCATGCAAGTCGAACGAGG-3') and An1388R (5'-
GCTTGACGGGCGGTGTG-3') were designed for the capture of greater phylogenetic

variability with 1380 bp amplicons.

DNA Extraction

DNA was isolated using the MoBio UltracleanTM Soil DNA kit (MoBio
Laboratories Inc.) and PCR amplifications of 16S rRNA genes targeting the anammox
group were performed in a model 9600 thermal cycler (Perkin-Elmer Cetus) from 30-80
ng uL'1 environmental DNA extracts. Reaction conditions for the primer pair Brod541F-
Brod1260R were: 10 pmol of each primer, 200 ng uL'l of bovine serum albumin (Roche),
2.5 U of Taq polymerase (Promega), 200 uM each deoxyribonucleoside triphosphate
(Invitrogen), 150 mM MgCl2 (Promega), and l/ 10 volume of 10X PCR buffer provided

with the enzyme. The PCR was as follows: 95°C for 3 min; 30 cycles of 95°C for 45 s,

25

60°C for 1 min for Brod541F-Brod1260R or 63°C for 1 min for An7F-An1388R, 72°C
for 1 min; and 72°C for 7 min. Clone libraries were created using TOPO TA Cloning®
(Invitrogen) and clones sequenced using an ABI 3730 Genetic Analyzer or ABI Prism
3700 DNA analyzer (Applied Biosystems). Sequences were analyzed against GenBank
with BLAST and the Ribosomal Database Project-II (RDP) (Cole et al. 2005). The
presence of chimeric sequences was determined by using the CHIMERA_CHECK
program from RDP. Sequences were aligned with ClustalW and phylogenetic trees
constructed using MEGA version 3.1 (Kumar et al. 2004) using a neighbor-joining

method with bootstrapping.

RESULTS AND DISCUSSION

The majority of primer sets designed for the amplification of anammox bacteria
did not yield consistent results. However, the screening primer set Brod541F-Brod126OR
resulted in consistent amplification in all sediments tested and was subsequently used to
create clone libraries of each site. The resulting 475 sequences exhibited greater than
96% nucleotide identity to either "Ca. Scalindua wagneri" or "Ca. Scalindua brodae," and
represented 100% specificity to the anammox group. Partial and redundant sequences
were removed and 698 bp fragments were aligned, resulting in 287 unique sequences.
These were compared against the anammox group consisting of: "Ca. Scalindua," ”Ca.
Brocadia," "Ca. Kuenenia," and "Ca. Jettenia asiatica" in addition to environmental and
wastewater obtained sequences that fall within the anammox group. A neighbor-joining

tree was constructed using the 25 sequences that remained after 98% redundancy was

26

removed (Figure 2.2) which excluded East Hanna Shoal Deep, Siberian Permafrost,
andWashington Margin sequences. Clones generated from Sherriff‘s Marsh and Siberian
permafrost exhibited the highest (88%) and the lowest (28%) percentage of unique
sequences, respectively. There was no discernible grouping either between freshwater
and marine samples (Figure 2.3). Thus, Brod541F-Brod1260R generated 16S rRNA
gene sequences from the eleven different sites cluster independently although sequences
from individual sites cluster together when 99% redundancy is included, reflecting a

certain level of endemism.

Additional clone libraries were created to confirm that the 698 bp Brod541F-
Brod1260R 16S rRNA gene sequences obtained did indeed represent full length
sequences of known anammox. Primer An7F, targeting "Ca. Scalindua," "Ca. Brocadia,"
and "Ca. Kuenenia," was first combined with the universal primer 1392R. The resulting
68 clones lacked any relationship to the anammox group, as was the case with the
Brod541F-1392R primer set, despite the increased specificity of the forward primers. To
resolve this, An7F was combined with An1388R and two 1380 bp sequences were
obtained out of 21 passing sequences. Amplification resulting from the An7F/1388R
primer set is illustrated in Figure 2.4. Clone FLE12 (positions 7-1388) was 97%
homologous to "Ca. Scalindua brodae" and 92% homologous to "Ca. Scalindua
wagneri." Positions 541-560 and 1241-1260 were identical to the Brod541F and
Brod1260R primer target sites, respectively. Clone FLC09 (positions 5-1388) was 96%

homologous to "Ca. Scalindua brodae" and 92% homologous to "Ca. Scalindua

27

Clone SHE A02
(HoneVVDfllH¥I
Clone SHE A08
Clone SHE C07
Clone WIN GlO
Clone EHS A04
Clone WIN E10
Clone WIN 012
Clone TURN E02
Clone TURN E04
Clone U3 H12 , . ..
Candidatus "Scalindua sorokmii"
(Candkhflus"Scahnduaqunnrf'
Clone EHS C04
TAF424468
Clone EHS F01
Clone Fl D10
Clone F 1 D3
Clone BUD H03
Clone U3 H04
Clone FUCA BIO
Clone U3 GOZ
lonelHJRIJIKXS
lone F1 D2
Clone F1 D7
Clone SHE B09
—* Clone SHE BIZ
Clone WIN E01
Clone EHS F07
r_—c ”33.91%; ttenia at ..
a tr e as a
(Sandkhuus"Ihmnnnnasunqpnfiensb"
(Jandthuus"Inocadrranannnoxuhnnw
AY769988
AY360085

AY266450

  
  
 
 
  
 
 
 
  

 

 

 

 

 

 

 

 

 

AY266449

 

 

AY360082

 

0.05

Figure 2.1. 168 rRNA gene-based bootstrapped phylogenetic tree reflecting the
relationship of Brod54lF-Brod1260R clones from East Hanna Shoal (EHS), East
Hanna Shoal Deep (Deep), Washington margin (Wash), Turning Basin (T urn),
Shallow Bud Inlet (Bud), Juan de Fuca (Fuca), Wintergreen Lake (Win), Sherrifi‘s
Marsh (She), Everglades F1 (F1), Everglades U3 (U3), and Siberian permafrost
(Perm) sediments with 98% redundancy removed with the anammox group and
related bacteria given as GenBank accession numbers. The bar indicates a 5%
sequence divergence.

28

l‘ PERM D03
'— SHE A08
. FUCA C02
'FUCA C06
‘WINH07

rFUCA BOZ
BUD E06

FUCA D04

TURN F03
FUCA D03

WIN E04
MICH SED (2)
SHE CO7

fSHE C02
DEEP C02

-- PERM C04
.. TURN E05

I . BUD (301
. TURN F09

—a EAST HANNA DEEP (2)
4 MICH SED (2)

iSHE D11
-‘E SHE B05

PERM A04

l——WINH06

SHE C11
BUD 010
EAST HANNA (3)

 

 

 

 

 

 

29

BUD c103
BUD F09
SHE Bos
FUCA C09
P4 MICH SED (5)
”’1 FUCA B08
SHE C08
PERM E04
BUD H11
DEEP D10
DEEP B12
P-——- SHE B04

 

PERM D07
“{ BUD F07

SHE C01

"l. DEEP C05
‘ FUCA BOS

r"- BUD G07
r'r- DEEP C05
_ DEEP BOS
'" TURN G04
a " BUD E01
'" SHE C10
l" BUD G06
BUD £02
' FUCAAIO

 

 

 

 

‘ MICH SED(3)
\

\ _

Figure 2.2. Neighbor-joining tree based on the 54lF/1260R primer set derived 16S
rRNA gene sequences after identical sequences were removed, illustrating the lack
of phylogenetic distinctiveness between freshwater and marine samples. Right to
left, top to bottom, with double bars indicating cutoff to the next tree. Groupings of
Michigan sediments and Everglades samples are noted by triangles. DEEP=East
Hanna Shoal Deep, Balt=four clone sequences derived from amplification of
Baltimore Harbor sediments by Tal et al., 2005, Contractor=168 rRNA sequences
derived from a rotating biological contractor.

1]

3+: BUDFlO

WIN H11

{ DEEP D03
TURN F06
- PERM E06
— BUD E03
'- DEEP All
‘LDEEP A04
TURN E09
"" DEEP D12
" F UCA B11
’ F UCA A08

 

 

 

Figure 2.2 cont.

' FUCA D02

" {- DEEP D09

WIN F02

. BUD G08

DEEP C11
*- PERM D12
_ PERM E05

1 BUD H09
f PERM D04

- BUD F04
- BUD E09
— DEEPAIO

‘ BUD H04
_[ DEEP B02
PERM F08

 

4 MlCH SED (3)

 

 

 

 

 

 

 

30

_ BUD H01
— BUD Fll
A DEEP B11

FUCA D06
WIN G02
WIN G01

— BUD F06
_ BUD F12

- WIN F06
E DEEP B09

*— PERM F04
FUCA B06
._ DEEP B01
TURN G12

— FUCA A09
— DEEP A09

EHS B02
BUD G09

 

* ——BUDF01

_ DEEP A08

— DEEP D05
F UCA D01

E DEEP B05
— TURN E07

 

meom
TURN E03
DEEP A02
MlCH SED (2)
MICH SED (2)

‘J.\

DEEP D08
BUD 611
BUD 10
FUCA A05
["- PERM F07
“r“ SHE 811
—4 MICH SED (5)
FUCA 12
BUD E05
BUD E11
WIN H01

r- FUCA C07

fl MICH SED (6)
TWIN 604

L- FUCA A02

‘ BUD H08

.WIN F11

- - SHE A11
E PERM F05

TURN F11
SHE D01
FUCA D08
TURN F08
TURN GOS
DEEP A12 .
J‘ MICH SED (5)

H SHE 012
SHE C12
,_. PER M06

MICH SE0 (4)
DEEP A03

3;“ BUD El 2

'i«—-—- WIN E10

 

 

 

 

 

 

 

  

 

 

Figure 2.2 cont.

31

FUCA C10
SHE C09
FUCA C05

SHE CO3
" DEEP D06

BUD H05
’ DEEP C06
f FUCA BO7
FUCA A04

““" DEEPC07
r" DEEP A07
. - SHE A06

 

, TURN E11

DEEP 806
PERM F06
BUD 612
TURN F10
WIN F08

'- BUD GOS
- DEEP DO4

 

 

r WIN E07

... FUCA 303
~ BUD H10

. DEEP c09

~ DEEP A05
, TURN F12
... DEEP 307
h—TURN E12

 

 

 

 

7d MICH SED (4)

4'!— EVERGLADES U3 H10
”TURN P04

_. EAST HANNA SHOAL (3)
__ TURN E04
,4 EVERGLADED U3 (6)
i~< EAST HANNA SHOAL (3)
. 5- EHs 305

J— EAST HANNA SHOAL (3)
23 L— MERTZ 1
,_ SCALINDUA WAGNERI
, SCALINDUA BRODAE
1 :___ SCALINDUA SOROKINII
l ‘“ SHE D04
ff EHS E11
.4 ‘“ JUAN DE FUCA (2)
TURN (308

U3 G05
MICH SED (5)
7‘" F1 010
:- EVERGLADES (2)
EHS B10
Q.»- EHS E08
~MERTz2
EHS P02
EHS D12
EHS C05
F1 D11
EHS F10
EHS A03
:EHS D07
' 3 EHS E09
f... EAST HANNA SHOAL (2)
. _ EHS D10

._ U3 (307
fr EVERGLADES U3 (2)

 

 

 

-1

..._..... n

 

 

 

 

 

Figure 2.2 cont.

32

TURN (302

——-U3 H04

. ”EVERGLADES (4)

jr-‘EVERGLADES (2)
*- F1 D12

'- PUCA D07

I P1 D5

}« EVERGLADES (2)
l— FUCA C04

L TURN (309

 

 

 

 

 

 

 

 

 

 

_{:31HUQE08
BLH)H03

, TIHUNF07

TIHUNED6

EVERGLADES (6)

"-—- FUCA BIO

DEEPSEA

JETTENLA

] KUENENIA

 

 

 

 

 

' CONTRACTORZ
AQUACULTURE

_____‘LSLUDGE
ANAMMOXIDANS

L CONTRACTOR

 

 

 

 

 

BALT 4

 

 

Figure 2.2 cont.

0.05 0.05

33

BALT l

BALT 2
BALT 3

I
'4 Iboeew-0

 

Figure 2.3. Amplicons resulting from the An7F/1388R primer set in Shallow Bud
Inlet, East Hanna Shoal, and Washington Margin samples. lkb ladders were used.

wagneri.” There was one mismatch within the Brod541F and 100% homology in the
Brod1260R site. A lack of homology to the Planctomycetes specific primer S-P-Planc-
0046-a-A-18 (Pla46F) (Neef et al. 1998) was observed in both clones. Sequencing error
was estimated at 4 bases in 1388 and anammox sequence targeting efficiency of this
primer was estimated to be 10% of the total retrieved passing sequences. A neighbor
joining tree was constructed using these full length sequences for comparison to the

anammox related clones derived from GenBank (Figure 2.5).

Specific 16S rRNA sequences belonging to the anammox group have historically
been difficult to recover from environmental samples. The Planctomycetes specific
forward primer S-P-Planc-0046-a-a-l8 (Pla46F), utilized as a "capture" primer for all

anammox, has been used to construct clone libraries, generally in samples which show

34

evidence of anammox activity via 15 N isotope pairing studies. However, Pla46F under-
represents anammox abundance in various studies (Egli et al. 2001, Schmid et al. 2000,
2003). For example, in an enriched sample where anammox represented 40% of the
bacterial population, only 36% of Pla46F derived 16S rRNA clones sequenced belonged
to the anammox group (Schmid et al. 2000). In enviromnental samples this method
results in very low target sequence retrieval efficiencies, possibly due to inherent PCR
template to product ratio bias (P012 and Cavanaugh 1998) or the lack of Pla46F
specificity. For example, the lack of Pla46 primer annealing site homology with the 1380
bp clone FLE12 and clone FLC09 sequences in this study would have resulted in the

omission of these clones from a Pla46F generated library.

In the context of this study, the novel primer set Brod54lF-Brod1260R is an efficient
method of screening environmental samples for the presence of ”Ca. Scalindua"-like
anammox. A previously suggested primer, S-*-BS-820-a-A-22 (Schmid et al. 2005), that
targets "Ca. Scalindua wagneri" and "Ca. Scalindua sorokinii", appears to be specific for
anammox bacteria with up to three mismatches but is only homologous with
approximately 50% of the sequences recovered in this study. Conversely, primer S-*-
Amx-0820-a-A-22 (Amx820), specific for genera "Ca. Brocadia" and "Ca. Kuenenia,"
(Schmid et al. 2000) exhibited at least three mismatches in approximately 80% and four
mismatches in 50% of the sequences retrieved with the Brod54lF-Brodl260R primer set

which was expected due to our targeting of "Ca. Scalindua" bacteria. Lastly, primer S-*-

35

AY360082

t0 outgroups

AY266450

AY266449

 

AY360085

 

    
 

Candidatus AF424468
"Scalindua brodae" Candidatus "Scalindua sorokinii"

Clone E12
Clone C09

AF 424463

   
   
   

Candidatus "Jettenia asiatica"

Candidatus "Brocadia anammoxidans"
AB054007

AB015552 Candidatus "Scalindua wagneri"

Candidatus "Kuenenia stuttgartiensis"

 

0.05

Figure 2.4. 16S rRNA gene-based bootstrapped neighbor joining tree reflecting the
relationship of full length An7F-Anl388R generated amplicons with the anammox
group as well as other sequences derived from environmental samples given as
GenBank accession numbers. The bar indicates a 5% sequence divergence.

36

Scabr-1114-a—A-22 (Scabrl 1 14), specific for "Ca. Scalindua brodae" and recommended
for use as a reverse primer with Pla46F (Schmid et al. 2003), exhibited at least 1
mismatch in 50% of the retrieved sequences. This suggests that use of these primers
could result in the under-representation of anammox in environmental samples and

illustrates the difficulty in specifically retrieving anammox sequences.

The identification of sequences with >96% nucleotide identity to the "Ca.
Scalindua" anammox in all freshwater, marine, and permafrost samples tested constitute
the first broad range 168 rRNA gene based molecular investigation of anammox
distribution. The presence of anammox 16S rRNA gene sequences in anoxic deep sea
sediments, shallow perturbed marine sediments, shallow organic rich freshwater
sediments, periphyton dominated aerobic sediments, and ancient frozen permafrost
sediments suggests that anammox bacteria in sediment are not constrained by conditions
that appear to be unfavorable to their metabolism and thus may be more widely
distributed than previously thought. While the presence of anammox sequences does not
equate to in-situ anammox activity, they do serve to identify habitats for more intensive
study and as molecular markers for better tracking candidate populations for anaerobic
ammonium oxidation. The wide distribution of anammox sequences in sediments
suggests that the process may not only be a significant oceanic nitrogen sink but may also

be an important method of nitrogen removal from freshwater wetlands and lakes.

The GenBank accession numbers for the sequences reported here are DQ869865-

DQ870184 and DQ869384-DQ8693 85.

37

ACKNOWLEDGEMENTS

Thanks goes out to Dr. Susan Newman and the staff of the South Florida Water
Management District for Everglades sampling support, David Gilichinsky of the Russian
Academy of Sciences for providing the permafrost sample and information, Stephen
Hamilton for providing the information of the Kellogg Biological Station sites, and Amy

Burgin for additional sampling assistance.

The National Science Foundation (OCE 0647891), Biotechnology Investigations-
Ocean Margins Program, Office of Biological and Environmental Research Office of

Science, US. Department of Energy provided financial support for this research.

38

REFERENCES

Bowman J.P., and RD. McCuaig. 2003. Biodiversity, community structural shifts, and

biogeography of prokaryotes within Antarctic continental shelf sediment. Appl. Environ.
Microbiol. 69:2463-2483.

Cole J.R., Chai B., Farris R.J., Wang Q., Kulam S.A., McGarrell D.M., Garrity
G.M., and J. M. Tiedje. 1 January 2005. The Ribosomal Database Project (RDP-II):
sequences and tools for high-throughput rRNA analysis. Nucleic Acids Res. 33(Database
Issue):D294-D296. doi:10.1093/nar/gki038.

Dalsgaard T., Canfield D.E., Petersen J., Thamdrup B., and J. Acur‘ia-Gonzalez.
2003. N2 production by the anammox reaction in the anoxic water column of Golfo
Dulce, Costa Rica. 422:606-608.

Dalsgaard T., and B. Thamdrup. 2002. Factors controlling anaerobic ammonium
oxidation with nitrite in marine sediments. Appl. Environ. Microbiol. 68:3 802-3 808.

Devol A. H. 2003. Nitrogen cycle-solution to a marine mystery. Nature 422:575-576.

Egli K., Franger U., Alvarez P.J.J., Siegrist H.R., Van der Meer J.R., and A. J. B.
Zehnder. 2001. Enrichment and characterization of an anammox bacterium from a

rotating biological contractor treating ammonium-rich leachate. Arch. Microbiol.
175:198-207. '

Freitag T. E., and J. I. Prosser. 2003. Community structure of ammonia-oxidizing
bacteria within anoxic marine sediments. Appl. Environ. Microbiol. 69:1359-1371.

Kirkpatrick J., Oakley B., Fuchsman C., Srinivasan S., Staley J.T., and J.W.
Murray. 2006. Diversity and distribution of planctomycetes and related bacteria in the
suboxic zone of the Black Sea. Appl. Environ. Microbiol. 72:3079-3083.

Kuai L., and W. Verstraete. 1998. Ammonium removal by the oxygen-limited
autotrophic nitrification-denitrifiation system. Appl. Environ. Microbiol. 64:4500-4506.

Kumar S., Tamura K., and M. Nei. 2004. MEGA3: Integrated software for Molecular
Evolutionary Genetics Analysis and sequence alignment. Briefings in Bioinformatics.
5:150-163.

Kuypers M. M. M., Lavik G., Woebken D., Schmid M., Fuchs B.M., Amann R.,
Jargensen B.B., and M.S.M. Jetten. 2005. Massive nitrogen loss from the Benguela
upwelling system through anaerobic ammonium oxidation.

Kuypers M. M. M., Sliekers A.O., Lavik G., Schmid M., Jergensen B.B., Kuenen

J.G., Sinninghe Damsté J.J., Strous M., and M. S. M. Jetten. 2003. Anaerobic
ammomium oxidation by anammox bacteria in the Black Sea. Nature. 422:608-611.

39

Neef A., Amann R.I., Schlesner H., and K.-H. Schlelfer. 1998. Monitoring a
widespread bacterial group: in situ detection of Planctomycetes with 16S rRNA-targeted
probes. Microbiology 144:3257-3266.

Polz M.F., and CM. Cavanaugh. 1998. Bias in template-to-product ratios in
multitemplate PCR. Appl. Environ. Microbiol. 64:3724-3730.

Risgaard-Peterson, N., Meyer R.L., Schmid M., Jetten M.S.M, Enrich-Prast A.,
Rysgaard S., and N. P. Revsbech. 2004. Anaerobic ammonium oxidation in an estuarine
sediment. Aquat. Microb. Ecol. 36:293-304.

Schalk J., Oustad H., Kuenen J.G., and M. S. M. Jetten. 1998. The anaerobic
oxidation of hydrazine: a novel reaction in microbial nitrogen metabolism. FEMS
Microbiol. Lett. 158:61-67.

Schmid, M. C., Maas B., Dapena A., van de Pas-Schoonen K., van de Vossenberg J.,
Kartal B., van Niftrik L., Schmidt 1., Kuypers M., Revsbech N.P., Mendez R., Jetten
M.S.M, and M. Strous. 2005. Biomarkers for in situ detection of anaerobic ammonium
oxidizing (anammox) bacteria. Appl. Environ. Microbiol. 71:1677-1684.

Schmid M., Twachtmann U., Klein M., Strous M., Juretschko S., Jetten M.S.M,
Metzger J., Schlelfer K.-H., and M. Wagner. 2000. Molecular evidence for genus level

diversity of bacteria capable of catalyzing anaerobic ammonium oxidation. Syst. Appl.
Microbiol. 23:93-106.

Schmid M., Walsh K., Webb R., Rijpstra W.I.C., van de Pas-Schoonen K.T.,
Verbruggen M.J., Hill T., Moffett B., Fuerst J., Schouten S., Sinninge Damsté J.S.,
Harris J., Shaw P., Jetten M.S.M., and M. Strous. 2003. Candidatus"Scalindua
brodae," sp. nov., Candidatus"Scalindua wagneri," sp. nov.: two new species of
anaerobic ammonium oxidizing bacteria. Syst. Appl. Microbiol. 26:529-538.

Strous M., Kuenen J.G., and M.S.M. Jetten. 1999. Key physiology of anaerobic
ammonium oxidation. Appl. Environ. Microbiol. 65:3248-3250.

Tal Y., Watts J.E.M., and H.J. Schreier. 2005. Anaerobic ammonia-oxidizing bacteria
and related activity in Baltimore inner harbor sediment. Appl. Environ. Microbiol.
71:_1816-1821.

Trimmer M., Nicholls J.C., and B. Deflandre. 2003. Anaerobic ammonium oxidation

measured in sediments along the Thames estuary, United Kingdom. Appl. Environ.
Microbiol. 69:6447-6454.

Van de Graaf A., Mulder A., de Bruijn P., Jetten M.S.M., Robertson L., and J.

Kuenen. 1995. Anaerobic oxidation of ammonium is a biologically mediated process.
Appl. Environ. Microbiol. 61: 1246-125 1 .

40

Vishnivetskaya T., Kathariou S., McGrath J., Gilichinsky D., and J. M. Tiedje.
2000. Low-temperature recovery strategies for the isolation of bacteria from ancient
permafrost sediments. Extremophiles 4: 165-1 73.

41

CHAPTER III

QUANTITATIVE PCR AND ACTIVITY OF ANAEROBIC AMMONIUM-
OXIDIZING BACTERIA IN DEEP-PACIFIC OCEAN SEDIMENTS OF THE

CASCADIA BASIN

ABSTRACT

The importance of anaerobic ammonium oxidation and the abundance of the
bacteria responsible for this alternative nitrogen removal pathway were determined in
deep continental margin sediments. Anammox was investigated in these 2800-3100m
below water surface suboxic sediments using 15 N amendments coupled with porewater
profiles and quantitative PCR of the Scalindua sp. type anammox 16S rRNA gene.
Removal of NH4+ was indicated in the suboxic zone while anammox rates ranged from
0.065 to 1.7 nmol N mL'l wet sediment h'l. Estimated anammox abundances exhibited a
mean density of 2 x 106 copies g'1 wet sediment and were correlated to anammox N2
production (R2=0.93). Significant numbers were found in the anoxic zone below the
depth where NO3' and N02' were identified. The overall contribution of anammox to total
N2 production ranged between 12 and 51%, in contrast to the general perception that

anammox activity increases with depth below water surface.

42

INTRODUCTION

First discovered in a wastewater treatment plant (Mulder et al., 1995), the
anaerobic ammonium oxidation (anammox) process was originally identified in the
environment within continental shelf sediments (Thamdrup and Dalsgaard, 2002). The
anammox process has been shown to be significant in ocean N removal and observations
from oxygen minimal zones show that fixed N is lost exclusively through this pathway
(Kuypers et al., 2005; Hamersley et al., 2007). Marine sediments are thought to be
responsible for about half of the marine N2 production and anammox activity has been
detected in these environments worldwide. However, the range of reported anammox
importance varies greatly, with estimates of anammox ranging between 0 and 80 percent
of total N2 production (Thamdrup and Dalsgaard, 2002; Risgaard-Petersen et al., 2004;
EngstrOm et al., 2005). Anammox has been thought to be responsible for an increasing
percentage of total N2 production as water depth increases (Dalsgaard et al., 2005), with
the highest reported conuibution at 700m depth in Skagerrak, Denmark (Thamdrup and
Dalsgaard, 2002; EngstrOm et al., 2005). This is thought to be due to the lower
availability of reactive organic carbon in deeper sediments which would favor anammox
over denitrification (Thamdrup and Dalsgaard, 2002) and would lead to anammox

dominance at depths greater than 1000m (Dalsgaard et al., 2005).

The bacteria responsible for the anammox process reside within deep
phylogenetic branches of the order Planctomycetales and have not yet been isolated in
pure culture. 16S ribosomal ribonucleic acid (rRNA) gene targeted assays suggest that
anammox bacteria are widespread in the environment (Penton et al., 2006), although the

extent of their activity in many of these habitats is unknown. While the isotopic method

43

for anammox N2 detection is now widely used for quantifying anammox N2 production,
the enumeration of anammox bacteria in sediments is performed less often. The
knowledge of population depth distribution when accompanied by N2 production rates of
anammox bacteria allows for the calculation of cell-specific N2 production. This can be
further implemented to investigate correlations between anammox activity and sediment
biogeochemistry. Fluorescence in—Situ hybridization (FISH), targeting the 16S rRNA
gene, has been the preferred method for enumerating anammox in environmental and
wastewater samples (Schmid et al., 2005). FISH counts of anammox cells have ranged
from 600 to 22,000 cells ml’1 in marine water column samples (Kuypers et al., 2003,
2005; Schubert et al., 2006). Conversely, Schmid and colleagues (2007) found 2.1x106 to
1.4x108 anammox cells ml'1 in nine diverse anoxic marine sediments while ~300
anammox cells ml'1 were found in Baltimore Harbor sediments (Tal et al., 2005) using

the probe AMX820.

Quantitative polymerase chain reaction (Q-PCR) is a reliable indicator of bacterial
numbers of targeted groups and can be an indicator of their activity (Gri'rntzig et al.,
2001; Qiu et al., 2004; Schippers et al., 2006). Recently, the number of anammox bacteria
closely related to ‘Candidatus Brocadia anammoxidans’ was determined in enrichment
cultures from a rotating reactor using a SYBR green Q-PCR assay targeting the 16S
rRNA gene (Tsushima et al., 2007). Compared to FISH, the Q-PCR 98-well format
allows for high sample throughput and eliminates investigator bias introduced by the
microscopic counting technique. This method is not affected by background

hybridization and the presence of dense cell clusters (Third et al., 2001) that can

44

influence cell counts in sediments, nor is it dependent on FISH hybridization stringencies

that can vary between sediment types.

In this study we use 15 N isotope incubations to study benthic anammox activity at
eight Sites off the Washington/Oregon coast at depths of 2800-3100m We also
quantitatively compare the anammox process at these sites with the catalytic bacteria, a
particularly important goal for understanding the biogeochemical processes that influence
anammox activity in the environment. Specifically, we developed a Q-PCR assay specific
for anammox bacteria most closely related to ‘Candidatus Scalindua’ (Schmid et al.,

2003), the only currently recognized marine anammox bacteria.
MATERIALS AND METHODS
Site description and sediment sampling

Cascadia Basin is a sedimented basin enclosed within the North American
margin, the Blanco Fracture zone and Juan the Fuca ridge, located at approximately 43-
48°N and 125-130°W at an average depth of 2000-3000m. Sediment was collected at
eight stations in the Cascadia Basin during a RN. T.G. Thompson cruise (TN 198) in
August 2006 (Figure 3.1). A multi corer was used to collect sediment cores (10 cm
diameter) with a well-preserved sediment-water interface; on average the cores had a
sediment depth of 40 cm and an overlying water column of 60 cm. Collected cores were

stored on board at 4°C, in the dark.
Oxygen and porewater profiles

Three oxygen profiles were obtained from a smaller sediment core (5 cm diameter, 25 cm

45

length) that was sub-sampled from one of the original cores with a Clark type 02 sensor
(Diamond General Corp. Ann Arbor, United States of America (U .S.A.), type 737; Lohse
et al. 1996). Oxygen profiles were measured within 1 hour Of core retrieval and

immediately after core sub-sampling.

A sediment core from each station was sectioned at 0.5 cm increments (0-5 cm sediment
depth) and 1 cm increments (5-20 cm sediment depth) for pore water N02, N03' and
NH]. Each section was placed in a 60 mL centrifuge tube, centrifuged for 20 min (11000
g) and supernatant water was filtered through a 0.45 pm cellulose acetate filter. All
sediment slicing and centrifugation was performed at 4°C. Inorganic N samples were
analysed within 5 hours from sampling, on a semi-automatic analyzer according to
standard colorimetric procedures (Strickland and Parsons 1972). Porosity was estimated
from sediment collected from the same profile as the inorganic N. Collected sediment
was stored in pre-weighed sealed glass vials until the end of the cruise and porosity was
calculated from the weight of water loss after drying the sediment (50°C) to constant
weight. Samples for pore water Mn2+ analysis were stored frozen, acidified with HCl
immediately prior to analysis and concentration was determined using a Perkin Elmer

flame AAS.

Nitrate consumption and ammonium production rates in the sediment were
calculated by taking the slope based on the decreasing values from the maximum N03'
concentration or the slope based on the increasing values from the minimum NH4+

concentration. Flux was then calculated using Fick’s first law of diffusion, assuming a

46

 

 

MON ..

 

 

46 °N

 

 

42°N

 

 

 

 

130°W 128°W 126°W 124°W

Figure 3.1. Map of the eight station locations (SI-SS) in the Cascadia Basin, North
East Pacific Ocean.

47

steady state. The sediment diffusion coefficient (DW) used in these calculations was
estimated from the free-solution diffusion coefficient (D0 ; Li and Gregory 1974) and the

measured sediment porosity ((1); Boudreau 1997).
15N incubations and analysis

Sediment throughout the suboxic zone was sectioned into 2 cm slices and
homogenized. Approximately 5 mL sediment was placed in 12 mL Exetainer (Labco,
High Wycombe, United Kingdom), sealed and gently flushed with helium. Pre-
incubation followed at 4°C in the dark for varying lengths of time in order to remove the
residual sediment 02, N02, and NO;;': 30 hours for stations 1, 2, and 3, 50 hours for
station 5, and 65 hours for stations 4, 6, 7, and 8. When the sediment was too
consolidated for homogenization, 24-60 mL of overlying water collected from each
respective core was added, increasing the porosity by 2-9 percent. This was taken into

consideration when calculating the final N2 production rates.

After the pre-incubations were completed, the 15N isotopes were added (Figure 3.2).
Three treatments were performed in which 15NH4+, ISNH4+ + 14NO2' and 15N02' were
added by introducing 0.1 mL of a 2 mmol L'1 or 4 mmol L1 l5N solution to the Exetainer
through a rubber septum using a Hamilton syringe (final concentrations ~ 40 nmol
15NH4+ mL'l (wet sediment), ~ 80 nmol ISNH4+ + 40 nmol 14NO2' mL'l (wet sediment),
~ 40 nmol '5N02' mL'l (wet sediment). According to Trimmer et al. (2005), potential

anammox rates show no response to added nitrite at concentrations over 10 pmol L'l

N02'. A Helium flow cap was used to prevent headspace contamination with 02 or N2

48

     

Isotope Ratio GC-MS

_ 29
29/30N2 = 29N2 = Anammox N2
Ambrent NO3 or Anammox Denitrification = 30N

NO2 present possrble 2

Figure 3.2. Figure illustrating the experimental setup for the determination of
anammox activity.

during isotope addition. Incubations were stopped with 0.1 mL 7 M ZnCl2 at regular

intervals.

The l5NH4+ treatment was used as control to detect oxidation of ammonium in the

sediment without the addition of nitrite. If all of the oxidants are removed during pre-

29/30

incubation, the 15NH4+ treatment should not show production of N2. In the parallel

+ - . . . .
15NH4 + 14N02 treatment, 29N2 production suggests that the added nrtnte rs the

oxidizing agent for ammonium. The combination of these two treatments establishes
anammox activity. Lastly, potential anammox and denitrification rates were estimated
from the 29N2 and 30N2 production rates in the 15NO2'-only treatment according to
Thamdrup and Dalsgaard (2002). Concentrations of 29’30N2 were determined using a
Finnigan-Mat 253 isotope ratio mass spectrometer and calculated as excess above natural

concentrations.

49

DNA extraction

Extruded whole sediment cores were sliced into 0.5 cm sections and
approximately 5 g of stirred sediment was sub-sampled and stored at -20°C until analysis.
Genomic DNA from 0.5 g sediment was isolated using the MoBio Ultraclean soil DNA
kit (MoBio Laboratories, Inc.) and further purified by electrophoresis in 0.8% Cambrex
SeaPlaque low melting temperature agarose at 4°C. Genomic DNA bands were gel-
extracted, treated with gelase (Epicentre) per manufacturer’s instructions, and
concentrated using Microcon columns. DNA concentrations were determined by
absorbance at 260 nm. The number of genomes g'1 sediment was estimated using the

conversion of 980 Mbp pg'l DNA and an average genome size of 4.3 Mbp.
Primer-probe design and specificity

We were unable to identify applicable primer sequences for SYBR green
quantitative anammox determination due to the relatively small number of unique
homologous regions of the 16S rRNA gene for specific amplification of anammox
without primer dimer formation. Hence, the Taqman assay was used to overcome these
limitations. Two primers were designed for the detection of anammox l6S rRNA genes:
541F RT (5’-GAG CAC GTA GGT GGG TTT GTA AG-3’) and 616RRT (5’-CCT CCT
ACA CTC AAG ACT YGC AG-3’). Primer 541FRT was based On the previously
described primer Brod541F, found to be 100% specific for anammox 16S rRNA genes
found in various environmental samples including the Juan de Fuca Ridge and
Washington Margin, both in the Pacific Ocean (Penton et al., 2006). The Taqman probe

AnPrb (5’-CAG RTG TGA AAG CCT TCT GTT CAA CGG AAG-3’) was labeled 5’

50

FAM and 3’ Black Hole Quencher-l® (Integrated DNA Technologies, Coralville, Iowa).
Alignments of 475 16S rRNA gene sequences obtained with Brod541F and Brod1260R
as well as environmental clones related to anammox from GenBank were used to confirm
primer and probe complementary regions and determine where the use of degenerative
bases was necessary. The primers and probe were designed with the program
PrimerExpress (PE Applied Biosystems), and verified against GenBank for specificity:
540RT, 616RT, and AnPrb respectively showed one, two, and three mismatches against
‘Candidatus Scalindua wagneri ’ (AY254882), none, none, and two (due to extra A’s in
the GenBank 16S sequence, possibly due to extra base calls) mismatches with
‘Candr'datus Scalindua brodae’ (AY254883), two, zero, and zero mismatches with
‘Candidatus Scalindua sorokinii ’: (AY257l81), and generally all had zero mismatches

with the 475 clones from our previous study (Penton et al., 2006).
Quantitative PCR assays

Quantitative PCR reactions and optimizations were performed as described in the
Applied Biosystems protocol for Taqman assays. Final assay concentrations of 541FRT,
616RRT, and AnPrb were 300 nmol L", 900 nmol L", and 250 nmol L", respectively
with 1 ,uL of genomic DNA added per well and annealing and extension temperatures of
61°C. The increase in fluorescence emission was monitored during PCR amplification
using a 7700 Sequence Detector (PE Applied Biosystems). The threshold cycle (CT)
values obtained for each sample were compared to the standard curve to determine the
initial 16S rRNA gene copy number and were expressed as ,ug" environmental DNA and
normalized to g" wet sediment. Q—PCR assays were performed on 0.5-cm core sections

while anammox incubations were performed on 2-cm sections. Five laboratory replicates

51

were used per 0.5 cm core section. The relationship between molecular data and '5N

incubation data were based on means obtained from the 0.5-cm sections per 2-cm sample.
Q—PCR sensitivity, detection limit, and accuracy

Sensitivity determination and standard curve generation were performed using an
anammox l6S rRNA gene containing clone which exhibited 99% nucleotide identity to
the characterized anammox ‘Candidatus Scalindua wagneri ’. The clone was created
using primers Brod541F/Brod1260R from various environmental samples (Penton et al.,
2006) and dilutions ranging from 4 to 8x108 copies were prepared in individual reaction
tubes. A calibration curve was constructed to determine our ability to discriminate two—
fold differences in template concentrations and the lower detection limit with assays
performed in triplicate. In order to test for the presence of PCR inhibitors, 1x104 clone
16S rRNA gene copies were added to representative samples from each site. COpy
numbers, with and without the addition of 1x104 copies, were compared and the percent

recovery from each sample was determined.

RESULTS
Porewater O2 and dissolved inorganic nitrogen

. Porewater and NH4+ and N03' profiles from continental margin sediments in the
area of this study (Lamboum et al. 1996; Stump and Emerson 2001), the western
Mexican margin (Hartnett and Devol 2003), California margin (Reimers et al. 1992) and

Panama Basin (Aller et al. 1998) all Show a pattern that indicates ammonium removal in

52

the nitrate reduction zone followed by ammonium accumulation after nitrate is depleted.
Another indication of nitrogen removal in these sediments is the low N:P ratios that
decrease from 6.5 to <1 in the suboxic zone (5-20 cm; Station 5 and 6 in Stump and

Emerson 2001).

Among the eight stations, oxygen penetrated 0.5 to 1 cm into the surface of the
sediments. Nitrate penetration differed widely between sites, but were generally divided
into two groups, those close to shore (S1, S2, and S3; 2.5 cm), and those farther from
shore (S4, S7, S8; 7-10 cm) (Figures 3.3a and b).. The porewater Mn2+ distribution was
measured at stations 82, S4, S6, and S8. Patterns showed an increase in Mn2+ in the
middle of the nitrate reduction zone. Although both Mn02 reduction and denitrification
release NH4+ to the porewater, porewater profiles Show a net ammonium removal in the

suboxic zone followed by ammonium accumulation starting at the nitrate depletion zone.
'SN incubation experiments

There was no evidence of residual oxidized nitrogen species following depletion
during pre-incubation in stations S1, S2, S3, SS, S6, and S7 as determined by the ”NH;
only incubations. However, at stations S4 and S8 NO2' and N03' were present at the zero
time, identified by the detection of 29N2. Small amounts of 29N2 (0.5 pmol L")
accumulated after the addition of ”N114”: but remained constant throughout the length of
the incubation. This indicated that the initial NO3' / N02’ was the only oxidant present.
Significant 29N2 production was detected through the suboxic zone at all stations after the
addition of both 15NH4+ + MNO2' (Figure 3.4). In the incubations where '5NO2' was

added, both 29N2 and 30N2 were formed immediately. 15N-labelled N2 could only be

53

 

 

 

 

 

 

 

 

 

4 0o 1+5n’M9 4
‘ W '49?

 

 

 

 

 

Figure 3.3a. Pore water distribution of NI-14+(A), N02' (0) and N03' (0) at stations
Sl-S4. At S2 and S4, Mn2+ (A) profiles are also shown. Note that the scale for

ammonium and Mn2+ concentrations vary between stations.

54

—<>—N0 ,‘(uMl —o—3N(p,(uM)

0102030405001020 50

 

 

0 4 812482024

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3. 3b. Pore water distribution of NH4 (A), N02 (0) and N03 (0) at stations
SS-S8. At S6 and S8, Mn2+ (A) profiles are also shown. Note that the scale for
ammonium and an+ concentrations vary between stations.

55

detected in the presence of NO3' / N02" which suggest that it is the dominant oxidant for

anaerobic ammonium oxidation.

Total anammox contribution to total N2 production ranged between 12-51 percent
with an average of 38 3:99 percent. Absolute rates varied between 0.065-1.7 nmol N
mL" wet sediment h" (anammox) and 017—1 .8 nmol N mL" wet sediment h"
(denitrification) (Figures 3.5a and b). Both potential denitrification and anammox activity
decrease with distance from shore (Figure 3.6) with a stronger relationship with
anammox activity (R2=0.67) than denitrification (R2=0.3 8). One sediment depth showed
an anammox contribution of only 12 percent, due to a low anammox rate (0.065 nmol N
mL" wet sediment h") when compared to the 4-6 cm depth at S7 (0.25 nmol N mL" wet
sediment h") where the anammox contribution was 38 percent. If this value is excluded

then the average anammox contribution was 40 327.4 percent.
Quantitative PCR Assays

Using 475 sequences with >96% homology to the anammox ‘Candidatus
Scalindua wagneri ’ and ‘Candidatus Scalindua brodae ’ obtained from our previous study
(Penton et al., 2006), we modified the forward primer by extension and inclusion of
degenerative bases for use with the Taqman quantitative PCR system. The Q-PCR
targeting of these specific anammox was further confirmed by performing GS FLX
pyrosequencing (Genome Sequencer FLX System, 454 Life Sciences, Inc., Bradford, CT,
USA) in station 2 surficial (0-2 cm) sediments using primers covering the 168 rRNA
gene V4 hypervariable region. Using the RDP Classifier to sort sequences into the

unidentified Planctomycetes, we identified 25 anammox sequences out of 41 18, all of

56

$9
.5
E
+

 

 

 

 

 

 

i .

'1 ‘ +30 2
2.

3 . ”2

g0 31 , LL
0 25 50 75
15-

E 12. “II-l,“ + 1“NO;

'- ~ 0

g 9 O o

E g o 0

'75 6‘

i 3;

E 0F————‘ 11 ,n._...n___,
0 25 5O 75
15-1 _

€12: m2 2 0 o
‘ O

P 9; . 0 e

5 6 °

i“:
011 . . . - .
O 25 50 75

T1me(hours)

Figure. 3.4. Production of 29N2 and 3”N2 over time in anoxic sediment incubations
following additions of 40 nmol lsNH4+ ml" (wet sediment), 80 nmol 15NH4+ + 40 nmol
l"NO2' ml" (wet sediment) and 40 nmol '5NO2' ml" (wet sediment); n=2. Data are
from 15N incubations at station SS, anammox and denitrification rates were
estimated from 29N 2 and 3"’N2 production rates in the 15N02' treatment.

57

- Denitrification

l:l Anammox 1 1
nmol N ml' wet sediment h'

0.5 1.0 1.5 2.0

 

 

 

Depth (cm)

  

$2

 

 

 

 

    
  

 

 

 

1 - A - - . - g A -
O 10 20 30 40 10 20 30 40
N03' (uM) NO3' (uM)
nmol N ml'l wet sediment h'1
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
- . . . - . . . - . . . . . Q .
W " /./.
2 ii 51% //./
E 4 j 1, ,e’.
I Be 29%
‘5. 6 . /
8 , if
3 ' ‘ 47%
10_As:1S4
10 20 30 40 ’ 10 20 30 4O 50
N03- (uM) N03- (uM)

Figure 3.5a. Absolute anammox (white bar) and denitrification (black bar) rates
estimated from anoxic l5N sediment incubations at stations Sl-S4. Error bars denote
standard deviation at the stations where a replicate core could be tested (S4).
Percent values show anammox contribution to total N2 production at each depth
where N2 production rates were measured. Pore water nitrate distribution (e)
indicate the suboxic zone at each site. In this study the suboxic zone is defined as the
zone below oxygen penetration depth where nitrate is present.

58

- Denitrification

 

 

 

 

l:lAnammox
nmol N ml'l wet sediment h"1
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
_ _ l w/ ./
A2.—~/ 311/
5,, ' .33.;
£5
a“? 6 36%

   

 

 

 

 

 

 

 

 

10 ‘10‘20_‘3o‘40 ‘10‘20_‘30 40
N03 (uM) NO3 (uM)
nmol N ml"1 wet sediment h'1
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
T T ' ' ' 'Pfi ' ' ' 7 I T '5—1
t1 ./. o /./.
A 2w“ H“ /./‘
E .. / » /'
\o/ O/. __—" O/. o
4: 4' / fl 3141
8. g 38% 1* ./'
O
o 6:1 ./ ‘
11. . ./
8<’/ ‘./
to“- -9 -37 . - $8
10 20_ 30 40 10 20_ 30 40
N03 (uM) N03 (uM)

Figure 3.5b. Absolute anammox (white bar) and denitrification (black bar) rates
estimated from anoxic lsN sediment incubations at stations S5-S8. Error bars denote
standard deviation at the stations where a replicate core could be tested (S5 and S8).
Percent values show anammox contribution to total N2 production at each depth
where N2 production rates were measured. Pore water nitrate distribution (0)
indicate the suboxic zone at each site. In this study the suboxic zone is defined as the
zone below oxygen penetration depth where nitrate is present.

59

I 3113m

 

 

308an
1.64 ‘ 3‘ . . .
‘3: $2 Denitnfiestion
*3: 12‘ $3.“ 2738m 2804er
g 33 . S7
is 08- 3° 2838'" 2730'"
2 SB 3°
04- 2922'"
E t.
0.0 - r . . . . . . a .
0 20 40 60 80 100
,7: 1.6: g,
- - . Anammox
21.2- .
3
IE 084 I :5
Z 83
E 041 $75 :4 S7
C I I
1 $8
0.0

 

 

02014078010100
Distanoefromshore(km)

Figure 3.6. Absolute denitrification (e) and anammox (O) rates in Cascadia Basin
distributed with distance from shore (km). Rates presented in the diagrams are
estimated from the top 2 cm of the suboxic zone (the zone below oxygen penetration
depth where nitrate is present). Distances were estimated with the GPS program
Tsunami 99 from Transas.

60

which were >98% homologous to the ‘Candidatus Scalindua’ group (Figure 3.7). No

other 168 rRNA sequences related to the anammox group were identified.

Target copy number was determined using standards containing a cloned full
length 16S rRNA gene sequences highly related (>98%) to Scalindua brodae.
Amplification was monitored by the increase in fluorescence per cycle and the Ct value,
the point at which fluorescence crosses the pre—determined threshold, was subsequently
determined (Figure 3.8). The Q-PCR assay sensitivity and detection limit was tested
using an anammox 16S rRNA gene clone diluted from 8x108 to 4 copies (Figure 3.9).
The calibration curve showed that fluorescence was detected as low as four copies,
however precision was limited at lower copy numbers in part due to increased error at
higher CT’s, similar to results from earlier studies (Grfintzig et al., 2001; Rodrigues and
Tiedje, 2006). A regression of mean threshold cycles (CT) and the template DNAs along
the dilution range from 100 (the determined detection limit) to 8x108 copies gave a R2 of
0.99. The accuracy of the system was then tested by the addition of 8x104 16S rRNA
gene copies to representative samples from each sediment core in the Q-PCR assay. The
PCR reaction was generally not inhibited, as ten of thirteen depths exhibited 90-100%
recovery with the exception of an unexplained 57% recovery at station 5 in the 2-4 cm

slice (Figure 3.10).

Anammox 16S rRN A genes were distributed through Cascadia Basin sediments at
a mean density of 2x106 16S rRNA gene copies ,ug'1 of total community DNA or 1.6x106
copies g'1 wet sediment and accounted for an estimated 0.1% to 5.2% of the total
bacterial genomes present (Figure 3.11) which ranged from 9x107 to 6x108 g'l wet

sediment. Using the 15N measurements from the accompanying paper the activity

61

measures were compared with anammox population densities in each 2 cm slice.
Anammox copy numbers, in yg'l DNA determined by Q-PCR correlated (R2=0.93) with
anammox N2 production (nmol N ml'1 wet sediment h'l) (Figure 3.12). This remained
after the two highest values were removed (R2=O.68) and when all values from each 0.5
cm slice were included (R2=0.65). Percent anammox copy number of total environmental
genomes and anammox N2 production were also linearly correlated (R2=0.94). Cell-
specific activities were calculated by dividing the anammox N2 production by anarrimox
168 rRNA gene copy number for each sample and ranged from 1.3 to 11.3 fmol
ammonium cell'l day]. There were no relationships between NO2', NO3', or NH4+
profiles and anammox bacterial abundance. A depth profile from the sediment surface to
22 cm depth at station 2 showed that anammox numbers remained relatively constant to a
depth of 12 cm, ranging from 9x105 to 4x106 g'1 and then decreased to 8.8x104 g'1 at 19

cm (Figure 3.13).
DISCUSSION

Anammox appears to be a significant pathway for sediment nitrogen removal in
Cascadia Basin with an average contribution of 40 percent. This is in contrast to the trend
of increasing relative anammox importance with increasing depth below water surface
(bws) (Thamdrup and Dalsgaard 2002; Engstrom et al. 2005). When this trend is
extrapolated to sediments greater than 1000m bws, anammox is suggested to dominate
over denitrification (Dalsgaard et al., 2005), in conflict with these findings. The
contribution of anammox of l2-51 percent is within the range of many shallow and

intermediate depth sediments (40-400 m) (Dalsgaard et al., 2005).

62

ETUWAV202H44UE
- ETUWAV202J1TRO
ETUWAV202H2L99
ETUWAV202G|SF 5
ETUWAV202G7X2X
ETUWAVZOZGRYSN
ETUWAVZOZGBL08
ETUWAV202H9208

ETUWAV202JXSGZ
ETUWAV202FZCZY

ETUWAV202H1 BPH

ETUWAV202FZP06
ETUWAV202|EGNY

ETUWAV202HYDNZ

ETUWAV202|TUV9
ETUWAV202HCZDZ

Scalindua wagneri

Scalindua brodae
'- Scalindua somkinii

— ETUWAV202HM7X6
- ETUWAV202FXV39
ETUWAV202| l E4X
ETUWAV2026N1GL
ETUWAV202HGY|6
ETUWAVZOZGDEQB
ETUWAV202|U8NH

ETUWAVZOZGWXMT

 

 

 

 

_[——- Kuenenia stuttgartiensis

Jettenia asiatica

 

 

 

 

 

 

 

—‘|: Brocadia fulgida
Brocadia anammoxidans
ETUWAV202F0526
Nitrospira sp.
0.1

Figure 3.7. Variable region 4 168 rRNA based UPGMA tree using sequences
obtained from 454 pyrosequencing in station 2 (0-2 cm depth) sediment.

63

 

Delta Rn

 

1e-1

1 e-2

1e-3

 

 

 

 

Figure 3.8. Output of SDS software for a standard curve using a clonal full length
Scalindua brodae 16S rRNA gene with replication.

64

 

35'-
'i'li
I
30‘ |
.1
2
0
>3 ..
U 25
E
o d
3
CD
0)
E 20‘
[—
15-
' l 1 T r l j l ' T j I ' l T I

 

1 2 3 4 5 6 7 8
Log anammox clone 16S rRNA gene copy number

Figure 3.9. Calibration curve linear regression of anammox l6S rRNA gene copy
numbers determined from a clone dilution series from 8 to 8x108 copies. CT is the
threshold cycle number.

65

100%'

80% ‘

60%‘

40%‘

20%

 

St1 St2 St4 St5 St6 St6 St7
0-2 2-4 3-5 0.5-2 1 -2.5 5-7 4-6
St2 St4 St4 St5 St6 St7

0-2 1-3 5-7 2-4 2.5-4 2-4

Figure 3.10. Percent recovery of 10" added 168 rRNA gene copies from each
sediment matrix.

66

E Copy # ug'l environmental DNA
- Copy # g'1 wet sediment
Total genomes g“1 wet sediment

109

108

107

10. *

105

103

102

10
0%“ iffy efieeeeeeee
’(RQ 161””:

C;
Q '1? 55% 5"); ‘o‘vfizvbfb 5"
er” 9“" 64‘" 9‘" G35" :9 23;? 92.09 631° :9 :9 653’

Log copy number
32

 

 

 

 

 

 

 

 

 

 

 

 

 

e «0

Figure 3.11. Q-PCR determined anammox copy number yg'l environmental
DNA, g'1 wet sediment, and total estimated genomes g" wet sediment at each station.

67

1.8‘

1.4"

1.0"

Anammox nmol N ml”l wet sediment h'l

 

 

 

I I I I I
5 6 7
Log anammox 16S rRNA gene copies ,ug" environmental DNA

Figure 3.12. Linear regression of Q-PCR determined anammox gene copy ,ug'l
environmental DNA versus anammox N2 production. Data reflects average copy
numbers per 2 cm slice used for the determination of N2 production

68

 

—.-— Log copy number g'| wet sediment

 

4 5 6 7
-.‘- N02- (flmols L-I)
1 2 3
l I ' F
‘ -------------- EM.“
/. "‘----—--------.“/
I ,A'” .
A
'i 5‘ /
,A .
9.
9
\ C
\
§ 0
t \\.
\
'9 ./

 

 

 

 

 

 

 

a

o/./ ’
l t T ' l w 1
20 40 60 80

—®— NH; (,umols L")

 

' T ' r
10 20 30
—I— No,- (,umols L")

Figure 3.13. Depth relationship between anammox copy number, NO2', NHX, and
N03' at station 2.

69

Absolute rates

Anammox and denitrification absolute rates were significantly lower (Students t-
test p=0.0002; p=0.015, respectively), at 0.065-1.7 nmol N mL'l wet sediment h'1 and
0.17-1.8 nmol N mL'l wet sediment h'1 respectively, than coastal sediments at 26-12
nmol N mL'1 wet sediment h'1 and 59-25 nmol N mL'l wet sediment h'1 respectively
(Hood Canal, U.S.A., 38-147 m; Tofino Bay, Canada, 70-113 m, and the Gullmar Fjord,
Sweden, 118 m). Two distinct groupings occurred with the stations closer to the
continental slope having the highest absolute rates (81, S2, and S3) while the more
refractive stations further from shore (S4, S6, S7, and 88) exhibited significantly lower (t-
test p=0.01) anammox and denitrification activity when the top 2 cm of the suboxic zone

were compared.

The decrease in anammox rates with distance from shore in this study (Figure 3.6)
suggests that anammox activity is dependent on organic matter mineralization for
ammonium supply. In these sediments ammonium was supplied by MnO2 reduction and
denitrification. As such, anammox N2 production should be limited by low NH4+
concentrations in more refractive deeper sea sediments, such as the mid ocean or abysmal
plain. Freshwater anammox bacteria have the potential to supply themselves with their
own NH4+ from nitrate reduction (Kartal et al., 2006, 2007), a process which has not been

confirmed in the marine species (Jensen et al., 2008).

A near-surface NH4+ peak was detected at all stations in this study from which
NH4+ initially decreased with depth (Fig. 3.3). This decrease in NH4+ concentrations can

be considered as a subsurface minimum in the suboxic zone due to anammox. However,

70

similar profiles have been observed by others who attributed them to an artifact induced
by centrifugation (Berelson et al. 1990; Hammond et al. 1996) or cell lysis of microbes
and meiofauna in the top surface sediment during decompression and warming when the
cores are brought up to surface (Aller et a1. 1998; Hall et al. 2007). In the Cascadia Basin
this subsurface peak was observed at all locations deeper than 2000 m (Fig. 3.3; Stump

and Emerson 2001 ).
Significance for the marine nitrogen budget

Denitrification rates in deep sea sediments are widely estimated using NO3' profiles
(Berelson et a1. 1990; Middelburg et a1. 1996; Brunnege‘ird et al. 2004). However, the
presence of anammox confounds this calculation. N2 production rates are calculated by
the downward nitrate flux, assuming the consumption of two mole of NO3' per mole N2
produced. However, anammox will consume one mole NO2' per mole of N2 produced,
due to the coupling with ammonium. As such, this method will potentially underestimate
total N2 production by 20 percent when an anammox contribution of 40 percent is

present, as in Cascadia Basin.
Quantitative PCR

Estimated anammox abundances of 1.9x105 to 4.8x106 g'1 wet sediment were one
to two orders of magnitude higher than previously reported for the water column using
FISH. Fluorescence microscopy techniques may underestimate total abundance,
especially in sediments where particle shielding and aggregates may result in lower
observed numbers. However, the quantitative PCR method may also be influenced by the

inherent problems associated with genomic DNA extraction and PCR biases. In

71

particular, FISH was used in a study that found ~300 anammox cells ml'1 sediment (Tal
et al., 2005) using the probe AMX820. Conversely, Schmid and colleagues (2007)
reported 2.1x106 to 1.4x108 anammox cells ml'1 sediment using the FISH probe B8820,

which is generally within one order of magnitude of that found in our study.

Anammox percent abundance was generally below 1% of the total estimated
bacterial community in our sediments and within the range found in the Black Sea anoxic
water column (0.75%), Benguela upwelling (~l%), and Lake Tanganyika (~1.4%) when
4,6-diamidino-2-phenylindo1e (DAPI) was used for total cell counts (Kuypers et al.,
2003, 2005; Schubert et al., 2006). Two exceptions were identified in the 0 to -2 cm
sediments from stations 1 and 2. Here anammox comprised 5.2% and 4.6% of the total
community, respectively, and was correlated to the highest two anammox N2 production
rates found (Engstrém et al., accompanying). The lack of significantly lower anammox
abundances in the sites located farther from shore that exhibited lower anammox
importance suggest that while anammox numbers remain relatively constant, their cell
specific metabolism is lower at these sites. Total estimated bacterial genomes were in
general agreement with 1x107 to 1x1010 cells g'l found in other marine sediments using
Q-PCR (Schippers and Neretin, 2006) and approximately a thousand times higher than
those reported in water columns (Schmidt et al., 1998). However, DNA extraction
efficiency and genome size can heavily influence the calculation of total genomes present

and, as such, these data should be regarded appropriately.

The linear relationship between anammox 16S rRN A gene copy numbers and 15N
production rates (Figure 3.12) suggest that the anammox bacteria we detected were

responsible for the anaerobic ammonium oxidation occurring in these sediments. A

72

similar relationship was found in the suboxic Namibian shelf waters (R2>0.8) when
fluorescence in-situ hybridization (FISH) was used for enumeration (Kuypers et al.,
2005). Likewise, the relationship between Q-PCR numbers of ‘Candidatus Brocadia
anammoxidans' related anammox bacteria and anammox N2 production was found
(R2=0.8l) in enrichment cultures from a wastewater-seeded laboratory reactor (Tsushima
et al., 2007). Cell-specific rates were comparable to those found in the Benguela
upwelling (4.5 fmol of ammonium cell'l day") (Kuypers et al., 2005), the Black Sea (3-4
fmol of ammonium cell'l day'l) (Kuypers et al., 2003), and Lake Tanganyika (18 frnol of
ammonium cell'l day'l) (Schubert et al., 2006) water columns, further validating our
results. However, a recent study using Q-PCR for anammox enumeration in laboratory
bioreactor enrichments found a low cell-specific activity of 0.072 frnol N cell'l day'l
(Tsushima et al., 2007), possibly due to Q-PCR amplification of dead or inactive cells
from substrate limiting conditions imposed by batch cultures. Cell-specific activities from
a recent marine sediment study (Schmid et al., 2007) showed activities that ranged
between 0.08 and 0.98 fmol N cell'l day", approximately a factor of ten lower than ours
despite our incubations at close to in-situ temperatures (4°C). A range of temperatures
appear to harbor similar cell specific activities, 15°C in marine sediments (Dalsgaard and
Thamdrup, 2002), 37°C in a wastewater treatment system (Strous et al., 2006), and our
activities at 4°C. This suggests that there are redundant enzymes optimized to different
temperatures or, alternatively, that there is a selective pressure on specific Scalindua
phylotypes which has been shown in enrichment cultures (Van de Vossenberg et al.,

2008)

73

The lack of relationship between anammox abundances and the N profile contrast
with the correlation found between anammox cell number and N02" in water columns
(Kuypers et al., 2003, 2005). Deficits of available NO2' and N03' below ~6 cm indicate
that significant anammox numbers are present where denitrification and anammox rates
should be negligible. A similar finding was reported in the Black Sea water column
(Kuypers et al., 2003) where anammox lipids did not decrease wholly in concert with
29N2 production and were present below where N02’ or NO3' was available. Anammox
presence in the deeper sediments could be due to alternative metabolic pathways evident
amongst other anammox lineages (Gfiven et al., 2005; Strous et al., 2006; Kartal et al.,
2007). This was recently verified in ‘Scalindua’ by van de Vossenberg (2008) and
colleagues who suggested that the widespread distribution may be due to metabolic
flexibility. Alternatively, sediment mixing by the significant benthic macrofauna
population may allow for transient NO3' / NO2' availability or the creation of preferential

porewater exchange channels for the maintenance of an anammox population.

Our results also show that Q-PCR based anammox 16S rRNA gene abundance in
these deep ocean sediments was correlated with 15N measurements and calculated cell-
specific activities were within the range reported previously. However, this correlation
significantly under-predicted measured anammox activity in productive Hood Canal
(Puget Sound, Washington, United States) sediments (P. Engstrom, unpubl.) and
significant copy numbers were also observed at depths where NO3' / NO2' was negligible.
As such, extrapolation of the relationship between anammox abundance and activities to
other locations other than those directly measured using 15N techniques is problematic.

Nonetheless, 16S rRNA-based Q-PCR provides insight into the enigmatic ecology of

74

these bacteria. Now that more is known about the metabolic flexibility of ‘Candidatus
Scalindua ’ we can begin to investigate the role they play in cycling at depths where N037
N02' is not available. Eventually, gene transcripts of the anammox pathway may serve as

a sensitive indicator of anammox activity.

The implementation of quantitative PCR and 15 N isotope pairing methods allowed
for a comparison between the catalyst and the process in these deep marine sediments.
The importance of anammox was lower than expected from several previous studies, and
did not dominate over denitrification in these continental shelf sediments. The lack of
correlation between anammox abundance and the N03' / NO2' profile suggests that other
controlling factors dominate the process and evidence from the porewater profiles

suggest an ammonium control.
ACKNOWLEDGEMENTS

Kind appreciation and recognition goes out to Dr. Pia Engstrc’im of the
Kristenberg Marine Station, University of Gothenberg, Sweden for the extensive
collaboration on this project. Dr. Allan Devol was responsible for the implementation of
this project and support. Professors P. Johnson, D. Hammond and S. Hautala are thanked
for their generous offer to let us join the cruise at the RN Thomas G Thomson. We also
acknowledge the helpful crew on R/V T.G. Thomson, M. Haught at the University of
Washington and J. C. Nicholls at Queen Mary, University of London for sharing their
mass spec expertise, and B. Chang at the University of Washington. The National

Science Foundation (OCE 0647891), Biotechnology Investigations-Ocean Margins

7S

Program, Office of Biological and Environmental Research Office of Science, US.

Department of Energy provided financial support for this research.

76

REFERENCES

Aller R.C., Hall P.O.J., Rude P.D., and J.Y. Aller. 1998. Bigoechemical heterogeneity
and suboxic diagenesis in hemipelagic sediments of the Panama basin. Deep-Sea Res. I
45: 133-165.

Berelson W.M., Hammond D.E., O'Neill D., Xu X.-m., Chin C. and J. Zukin. 1990.
Benthic fluxes and pore water studies fiom sediments of the central equatorial north
Pacific: Nutrient diagenesis. Geochim Cosmochim Acta 54: 3001-3012.

Boudreau B.P. 1997. Diagenetic models and their implementation. Springer.

Brunnegard J., Grandel S., Stahl H., Tengberg A. and P.O.J. Hall. 2004. Nitrogen
cycling in deep-sea sediments of the Porcupine Abyssal Plain, NE Atlantic.
Progr.Oceanogr. 63: 159-181.

Codispoti L. A. 2006. An oceanic fixed nitrogen sink exceeding 400 Tg N a-l vs the
concept of homeostasis in the fixed-nitrogen inventory. Biogeosci. Discuss. 3: 1203-
1246.

Dalsgaard T., and B. Thamdrup. 2002. Factors controlling anaerobic ammonium
oxidation with nitrite in marine sediments. Appl. Environ. Microbiol. 68:3802-3808.

Dalsgaard, T., Thamdrup B., and D.E. Canfield. 2005. Anaerobic ammonium
oxidation (anammox) in the marine environment. Res. Microbiol. 156: 457-464.

Engstriim, P., Dalsgaard T., Hulth S. and RC. Aller. 2005. Anaerobic ammonium
oxidation by nitrite (anammox): Implications for N2 production in coastal marine
sediments. Geochim. Cosmochim. Acta 69: 2057—2065.

Engstriim P., Penton C.R., and A.H. Devol. 2008. Anaerobic ammonium oxidation in
deep-sea sediments off the Washington Margin.

Griintzig V., Nold S.C., Zhou J., and J.M. Tiedje. 2001. Pseudomonas stutzeri nitrite
reductase gene abundance in environmental samples measured by real—time PCR. Appl.
Environ. Microbiol. 67:760-768.

Giiven D., Dapena A., Kartal B., Schmid M.C., Maas B., van de Pas-Schoonen K.,
Sozen S., Mendez R., Op den Camp H.J.M., Jetten M.S.M., Strous M., and I.
Schmidt. 2005. Propionate oxidation by and methanol inhibition of anaerobic ammonium
oxidizing bacteria. Appl Environ Microbiol 71: 1 066—1 071.

Hall P.O.J., Brunnegard J., Hulthe G., Martin W.R., Stahl H. and A. Tengberg.

2007. Dissolved organic matter in abyssal sediments: Core recovery artifacts. Limnol.
Oceanogr. 52: 19—31.

77

Hamersley M.R., Lavik G., Woebken D., Rattray J.E., Lam P., Hopmans E.C.,
Sinninghe Damsté J.S., Kriiger S., Graco M., Gutie’rrez D., and M.M.M. Kuypers.
2007. Anaerobic ammonium oxidation in the Peruvian oxygen minimum zone.
Limnol.Oceanogr. 52: 923-933.

Hammond D.E., McManus J., Berelson W.M., Kilgore T.E., and RH. Pope. 1996.
Early diagenesis of organic material in equatorial Pacific sediments: Stoichiometry and
kinetics. Deep-Sea Res. 11 43: 1365-1412.

Hartnett HE. and A.H. Devol. 2003. Role of a strong oxygen-deficient zone in the
preservation and degradation of organic matter: a carbon budget for the continental

margins of northwest Mexico and Washington State. Geochimic. Cosmochim. Acta 67:
247-264.

Jensen M.M., Kuypers M.M.M., Lavik G. and B. Thamdrup. 2008. Rates and
regulation of anaerobic ammonium oxidation and denitrification in the Black Sea.
Limnol. Oceanogr. 53: 23-36.

Kartal, B., Kuypers M.M.M., Lavik G., Schalk Huub J., Op den Camp J.M., Jetten
M.S.M, and M. Strous. 2006. Anammox bacteria disguised as denitrifiers: nitrate
reduction to dinitrogen gas via nitrite and ammonium. Environ. Microbiol. 9: 63 5-642.

Kartal B., Rattray J., van Niftrik L., van de Vossenberg J., Schmid M., Webb R.I.,
Schouten S., F uerst J.A., Sinninghe Damsté J., Jetten M.S.M, and M. Strous. 2007.
Candidatus ‘Anammoxoglobuspropionicus’ gen. nov., sp. nov., a new propionate

oxidizing species of anaerobic ammonium oxidizing bacteria. Syst. Appl. Microbiol.
30:39-49.

Kuypers M.M.M., Slieker A.O., Lavik G., Schmid M., Jorgensen B.B., Kuenen J.G.,
Sinninghe Damste J.S., Strous M., and M.S.M. Jetten. 2003. Anaerobic ammonium
oxidation by anammox bacteria in the Black Sea. Nature 422:608-611.

Kuypers M.M.M., Lavik G., Woebken D., Schmid M., Fuchs B.M., Amann R.,
Jorgensen B.B., and M.S.M. Jetten. 2005. Massive nitrogen loss from the Benguela
upwelling system through anaerobic ammonium oxidation. PNAS 102:6478-6483.

Lamboum L.D., Hartnett, H. and A.H. Devol. 1996. Special report NO. 113 Pore
water data from the Washington shelf and slope: Cruise WE9407B of the RN Wecoma,
University of Washington. 31 pp.

Li Y.H. and S. Gregory. 1974. Diffusion of ions in sea water and in deep-sea sediments.
Geochim. Cosmochim. Acta 38: 703-714.

78

Lohse, L., Kloosterhuis H.T., van Raaphorst W., and W. Helder. 1996.
Denitrification rates as measured by the isotope paring method and by the acetylene
inhibition technique in continental shelf sediments of the North Sea. Mar. Ecol. Progr.
Ser. 132: 169-179.

Middelburg J.J., Soetart K., Herman P.M.J., and C.H.R. Heip. 1996. Denitrification
in marine sediments: a model study. Glob. Biogeochem. Cycl. 10: 661-673.

Middelburg J.J., Soetart K., and P.M.J. Herman. 1997. Empirical relationships for
use in global diagenetic models. Deep-Sea Res. 1 44: 327-344.

Mulder A., Vandegraaf A.A., Robertson L.A., and J.G. Kuenen. 1995. Anaerobic
ammonium oxidation discovered in a denitrifying fluidized-bed reactor. FEMS Microbiol
Ecol 16:177—183.

Penton C.R., Devol A.H., and J.M. Tiedje. 2006. Molecular evidence for the broad
distribution of anaerobic ammonium-oxidizing bacteria in freshwater and marine
sediments. Appl. Environ. Microbiol. 72:6829-6832.

Qiu X.Y., Hurt R.A., Wu L.Y., Chen C.H., Tiedje J.M., and J.Z. Zhou. 2004.
Detection and quantification of copper-denitrifying bacteria by quantitative competitive
PCR. J. Microbiol. Methods 59: 199-210.

Reimers C.E., Jahnke R.A., and D.C. McCorkel. 1992. Carbon fluxes and burial rates
over the continental slope of California with implications for the global carbon cycle.
Glob. Biogeochem. Cycl. 6: 199-224.

Risgaard-Petersen N., Meyer R.L., Schmid M.C., Jetten M.S.M., Enrich-Prast A.,
Rysgaard S., and N.P. Revsbech. 2004. Anaerobic ammonium oxidation in an estuarine
sediment. Aquat. Microb. Ecol. 36: 293-304.

Rodrigues D.F. and J.M. Tiedje. 2006. Multi-locus real-time PCR for quantitation of
bacteria in the environment reveals Exiguobacterium to be prevalent in permafrost.
FEMS Microbiol. Ecol. 59:489-499.

Schippers A., and L.N. Neretin. 2006. Quantification of microbial communities in near-

surface and deeply buried marine sediments on the Peru continental margin using real-
time PCR. Environ. Microbiol. 8: 1251-1260.

Schmid M., Walsh K., Webb R., Rijpstra W.L.C., van de Pas-Schoonen K.T.,
Verbruggen M.J., Hill T., Moffett B., Fuerst J., Schouten S., Sinninghe Damsté, J.S.
Harris J., Shaw P., Jetten M.S.M, and M. Strous. 2003. Candidatus “Scalindua
brodae,” sp. nov., Candidatus “Scalindua wagneri”, sp. nov., two new species of
anaerobic ammonium oxidizing bacteria. Syst. Appl. Microbiol. 26:529—538.

79

Schmid M.C., Maas B., Dapena A., van de Pas-Schoonen K., van de Vossenberg J.,
Kartal B., van Niftrik L., Schmidt 1., Cirpus 1., Kuenen J.G., Wagner M., Sinninghe
Damsté J.S., Kuypers M., Revsbech P., Mendez R., Jetten M.S.M., and M. Strous.
2005. Biomarkers for the in situ detection of anaerobic ammonium-oxidizing (anammox)
bacteria. Appl. Environ. Microbiol. 71: 1677-1684.

Schmid M.C., Risgaard-Petersen N., van de Vossenberg J., Kuypers M.M.M., Lavik
G., Petersen J., Hulth S., Thamdrup B., Canfield D., Dalsgaard T., Rysgaard S., Sejr
M.K., Strous M., Op den Camp H.J.M., and M.S.M. Jetten. 2007. Anaerobic
ammonium-oxidizing bacteria in marine environments: widespread occurrence but low
diversity. Environ. Microbiol. 9: 1476-1484.

Schmidt J.L., Deming J.W., Jumars P.A., and RG. Keil. 1998. Constancy of bacterial
abundances in surficial marine sediments. Limnol. Oceanogr.43:976-982.

Schubert C.J., Durisch-Kaiser E., Wehrli B., Thamdrup B., Lam P., and M.M.
Kuypers. 2006. Anaerobic ammonium oxidation in a tropical freshwater system (Lake
Tanganyika). Environ. Microbiol. 8: 1857-1863.

Strickland J.D.H. and TR. Parsons. 1972. A practical handbook of sea water analysis.
Bull. Fish Res. Ed. Can. Bull. 169: 1-311.

Strous, M., Pelletier E., Mangenot S., Rattei T., Lehner A., Taylo M.W., Horn M.,
Daims H., Bartol-Mavel D., Wincker P., Barbe V., F onknechten N., Vallenet D.,
Segurens B., Schenowitz-Truong C., Médigue C., Collingro A., Snel B., Dutilh B.E.,
Op den Camp H.J.M., van der Drift C., Cirpus 1., van de Pas-Schoonen K.T.,
Harhangi H.R., van Niftrik L., Schmid M., Keltjens J., van de Vossenberg J., Kartal
B., Meier H., Frishman D., Huynen M.A., Mewes H.-W., Weissenbach J., Jetten
M.S.M., Wagner M., and D. Le Paslier. 2006. Deciphering the evolution and
metabolism of an anammox bacterium from a community genome. Nature 440:790—794.

Stump C. and S. Emerson 2001. Redox states of marine sediments from a transect along

47 N in the north Pacific Ocean; TTN 131 of the R/V T.G. Thompson. Special Report
No. 117, University of Washington. 57 pp.

Tal Y., Watts J.E.M., and H.J. Schreier. 2005. Anaerobic ammonia-oxidizing bacteria
and related activity in Baltimore inner harbor sediment. Appl. Environ. Microbiol.
71:1816-1821.

Thamdrup B., and T. Dalsgaard. 2002. Production of N2 through anaerobic ammonium

oxidation coupled to nitrate reduction in marine sediments. Appl. Environ. Microbiol.
68:1312-1318.

Thamdrup B., Dalsgaard T., Jensen M.M., Ulloa 0., Farias L., and R. Escribano.

2006. Anaerobic ammonium oxidation in the oxygen-deficient waters off northern Chile.
Limnol. Oceanogr. 51: 2145-2156.

80

Third K.A., Sliekers A.O., Kuenen J.G., and M.S.M. Jetten. 2001. The CANON
system (completely autotrophic nitrogen-removal over nitrite) under ammonium

limitation: interaction and competition between three groups of bacteria. Syst. Appl.
Microbiol. 24:588-596.

Trimmer M., Nicholls J.C., Morley N., Davies C.A., and J. Aldridge. 2005. Biphasic
behavior of anammox regulated by nitrite and nitrate in an estuarine sediment. Appl.
Environ. Microbiol. 4: 1923-1930.

Tsushima I., Kindaichi T., and S. Okabe. 2007. Quantification of anaerobic
ammonium-oxidizing bacteria in enrichment cultures by real-time PCR. Water Res.
41:785-794.

Van de Vossenberg J., Rattray J.E., Geerts W., Kartal B., van Niftrik L., van
Donselaar E.G., Sinninghe Damsté J.S., Strous M., and M.S.M. Jetten. 2008.
Enrichment and characterization of marine anammox bacteria associated with global
nitrogen gas production. Environ. Microbiol. In press.

81

CHAPTER IV

COMPARISON OF MICROBIAL COMMUNITY STRUCTURE

AMONG DIVERSE MARINE SEDIMENTS USING PYROSEQUENCING

ABSTRACT

Investigations of microbial community structure in environmental samples have
generally relied on the use of 16S rRNA gene based clone libraries. Sequences obtained
from these studies generally only reflect the most abundant members, making
comparisons that yield information about rare members difficult. We employed l6S
rRNA based pyrosequencing to investigate microbial community structure among ten
marine sediments. A total of 44,1 81 16S rRNA sequences were obtained that were placed
into 8031 clusters at a 95% similarity level. The top ten clusters accounted for 12.5% of
the total while 3882 clusters were represented only by one sequence in one sample. There
were no cosmopolitan clusters that were identified in every site. Cluster abundance based
trees showed that there was no relationship among sites based on geographic position,
depth below water column, or temperature. Our data indicates that 7.9% of the 44,181
sequences examined were identical to those recovered from cultivated bacteria,
challenging the notion that “less than 1% of bacteria are cultivated”. Of the total, 1456
sequences were highly divergent (<85% BLAST identity) from the public database with
1063 of these occurring in one cluster in an Arctic Barrow Canyon sample. The rare
sequences exhibited an overall lower identity to the database, which suggests that they

are an example of microbial endemism and not transients that would have a higher

82

probability of being sampled in another study. To examine the effect of nutrient loading
on similar sediments, we compared two Florida Bay sites that differ in phosphorus
availability. The relative abundance distribution of B-Proteobacteria, Actinobacteria,
Bacteroidetes, and S-Proteobacteria suggest that phosphorus controlled viral and grazer
predation may be playing a large role in shaping the microbial community structure in

these sites.

INTRODUCTION

Over the past 20 years culture independent methods have provided the
breakthroughs in attempts to catalog the vast diversity of microorganisms on Earth. A
biosphere has been unveiled, richer than imagined, with countless species whose
metabolic roles still defy description. Yet, we are still far from classifying the vast extent
of microbial diversity, as revealed by the myriad of rarefaction curves present in the
literature. A large portion of microbial diversity is due to rare species whom have
remained cloaked by dominant members due to the general lack of clone-based methods
and inherent cloning biases. Dominant species almost surely represent the overwhelming
functional component of an environment and are responsible for the highest proportion of
microbial biomass. The majority of microbes in nature are represented in relatively low
abundances (Pedros-Alio, 2006; Ashby et al., 2007; Elshahed et al., 2008) and have been
collectively coined the “rare biosphere” (Sogin etal., 2006).

Pyrosequencing, in which large surveys of microbial diversity are possible at a

relatively low cost, is a new technique that permits rarity to become visible. Overall, the

83

majority of rare sequences generated by pyrosequencing are more divergent from the
public database than the abundant members (Elshahed et al., 2008; Sogin et al., 2008)
with some phyla exhibiting a higher dissimilarity than others. Classical sequencing fails
to capture the members, even among the range of different environmental conditions.
Understanding this diversity is not trivial; it is the key to discovering novel genes, drugs,
and metabolic pathways. An example is the relatively recent discovery of the importance
of anammox bacteria in global nitrogen cycling (Thamdrup and Dalsgaard, 2002). Only
represented by a few sequences among numerous marine diversity assays, anammox
bacteria were typical rare members of the biosphere, as determined by their currently low
database representation. Through the use of targeted assays, they are now known to be
abundant at 108 cells g'l sediment (Schmid etal., 2007) and are responsible for up to 80%
of N loss in some locations (Thamdrup and Dalsgaard, 2002). In contrast, the majority of
rare species may exhibit little functional capability and not be relevant to the overall
carbon and nutrient flow.

One widely held notion is that “everything is everywhere, the environment
selects” (Baas Becking, 1934). Among the countless environments that have been
sampled, the marine biosphere harbors some of the highest microbial diversity on Earth
(Lozupone and Knight, 2007). Despite numerous studies, even the magnitude of
prokaryotic diversity in i the world’s oceans remains unknown and subject to debate
(Whitfield, 2005). Estimates have ranged from a few thousand taxa (Hagstrtim et al.,
2002), 103-104 taxa (Pommier et al., 2005), 106 taxa (Curtis et al., 2002), to higher
numbers estimated using rarefaction curves generated from 16S rRNA gene targeted

pyrosequencing assays (Sogin et al., 2006).

84

The persistence of species due to predation and competition in the marine
environment may be important factors that influence overall diversity and the rare
biosphere. Abundant members are known to be more susceptible to viral infection due to
simple encounter probability (Thingstad, 2000) and grazers prey selectively on the largest
and most active bacteria (Pemthaler, 2005). Environments that support higher numbers of
bacterial grazers or viruses may be expected to harbor a higher diversity and more rare
members, when all else remains equal. As such, there may be a survival strategy for the
rare members that are less susceptible to competition and predation (Sogin et al., 2006)
which otherwise may not be able to survive amongst more dominant members.

Here we report an assessment of microbial community diversity and structure
among ten marine sediments using a pyrosequencing approach. Our primers target the V4
16S rRNA gene hypervariable region, which has been shown to provide high coverage,
resolution, and accuracy for deep taxa classification. We further investigated the
relationship among two contrasting sites representing high and low phosphorus levels in

Florida Bay to test the effect of nutrient availability on diversity patterns.

MATERIALS AND METHODS
Sediment Sampling
Marine sediment samples were collected from ten sites (Figure 4.1) and the top 2
cm of sediment was used for DNA extraction unless otherwise noted: East Hanna Shoal
(EHS, 160 m depth, 72.637N 158.667W) and Barrow Canyon (BC, 186 m depth,
71.607N 156.214W), both from the Alaskan maritime in the Chuckchi Sea; Washington

Margin (WM, 1138 m depth, 46.575N 124.822W), West of the Juan de Fuca Ridge (JF,

85

Washington Margin

.H. f. 1
:5; f (00);“
. '. . Ii .,. I ‘

Gulfof Mexico . .
1 and 2 '

. ..
.. -
1 ' ,1
a.

Florida
Florida Ba 11

‘ L r- tlthlc
I ‘ “‘ m ‘

 

Figure 4.1. Sampling site locations in the Arctic and Pacific Ocean (top) and in the
Gulf of Mexico and Florida Bay (lower). The inset shows the location of the Florida
Bay sites west of the Florida Keys. Maps generated with GoogleTM Earth
(earth.google.com). Inset from South Florida Water Management District Landsat 7
ETM 2000 Florida Coastal Everglades LTER Mapserver Project.

86

3869 m depth, 46.783N 133.667W), and Cascadia Basin (CB, 42.000N, 125.230 W) all
in the Pacific Ocean; two samples from one site (800 m depth, 26.404N 96.064W) in the
Gulf of Mexico (GM 1, 0-0.5 cm) and (GM2, 4-5 cm); 3 samples from shallow marine
habitats, Florida Bay 9 (FL9, 25.177N 80.490W), Florida Bay 10 (FL10, 25.025N
80.681W), and Florida Bay 11 (FLl 1, 24.913N 80.938W). DNA was isolated using the
MoBio UltracleanTM Soil DNA kit (MoBio Laboratories Inc.) according to the

manufacturer’s instructions.

SSU rRNA Amplicon Pyrosequencing

Eubacterial universal primers were used that spanned the SSU rRNA gene v4
hypervariable region, corresponding to Escherichia coli SSU rRNA positions 563 to 802.
Primer coverage was determined by the internal alignment of perfect matches against
SSU rRNA gene sequences in the Ribosomal Database Project II (RDP) (93.6%
coverage) and from the Sorcerer 11 Global Ocean Sampling Expedition (94.6% coverage)
(Rusch et al., 2007).

The forward key-tagged primers were composed of sequencing adaptor A,
sample-specific 4 or 6-bp keys, and a eubacterial 563F primer (5’-
GCCTCCCTCGCGCCATCAG(keys)AYTGGGYDTAAAGVG-3’). The reverse fusion
primer consisted of sequencing adaptor B, and a eubacterial 802R primer (5’-
GCCTTGCCAGCCCGCTCAGTACNVGGGTATCTAATCC-3’). All primers were
dual HPLC-purified (Integrated DNA Technologies, Coralville, IA) to increase
specificity of primers and minimize the miss-sorting of samples by primer synthesis

error. PCR mixtures contained 1 nM of each primers (IDT, Coralville, IA), 1.8 mM

87

MgC12, 0.2 M dNTPs, 1.5 X BSA (New England Biolabs, Beverly, MA), 1 unit of
FastStart High Fidelity PCR System enzyme blend (Roche Applied Science, Indianapolis,
IN), and 10 ng of DNA template. Amplification conditions were 30 cycles of 95°C for 45
sec, 57°C for 45 sec, and 72°C for 1 min after an initial denaturation for 3 min at 95°C,
followed a final 4 min extension at 72°C. For each sample, three replicated PCR reactions
were run in parallel and excised bands in range of 270-300 bps were combined.
Amplicon recovery was performed with Qiagen Gel extraction (Qiagen, Valencia, CA)
followed by an extra Qiagen PCR Purification step. DNA was Quantified
spectrophotometrically using the NanoDrop ND-1000 spectrophotometer (NanoDrop
Technologies, Wilmington, DE) and equimolar amounts of each sample were
subsequently combined and subjected to pyrosequencing using Genome Sequencer F LX

System (454 Life Sciences).

Pyrosequencing Analysis

Raw reads were sorted into individual samples using the assigned key sequence
and filtered using two categories of sufficient read length and a less than 2bp primer edit
distance. Sequences were aligned using INFERNAL version 8.1, a stochastic context-free
grammar based aligner (http://infemal.ianelia.org/) along with the Bacterial SSU rRNA
secondary-structure model of Gutell and colleagues (Cannone et al., 2002). All sequences
were then clustered by the complete-linkage method at the desired distance and assigned
to bacterial taxa using the RDP Classifier trained on species-type strain sequences from
the Taxonomic Outline of Bacteria (TOBA; http://wvov.taxonomicoutline.org) along with

additional sequences for regions of bacterial diversity not covered well by TOBA. The

88

bootstrap confidence threshold was set to 50%. Abundance-based adjusted Jaccard and
Sorensen indices (Chao et al., 2006) were calculated for each pair of samples using

Estimates (http://viceroyeeb.uconn.edu/estimates) after clustering samples together.

RESULTS

Phylogenetic Structure

We sequenced 44,181 SSU rRNA gene V4 hypervariable region bacterial
amplicons from ten sites. Operational taxonomic units (OTUs) were clustered at the 95%
similarity level as this corresponds to the diversity observed between species type-strains
from the same genus. These 8031 clusters were used for the generation of rarefaction
curves and the calculation of Shannon diversity and Chaol species richness estimates
(Table 4.1). The highest diversity and richness estimate was found in sample F L11 while
the lowest values were associated with Barrow Canyon. The difference in estimated
diversity between these two contrasting sites was not illustrated in large scale changes in
relative abundances among the most prolific phyla (Figure 4.2). The phylogenetic
architecture (Acinas et al., 2004) indicates that there was extensive deep-lineage variation
among sites (Figure 4.3).

The top ten clusters accounted for 5501 sequences or 12.5% of the total while
3882 clusters (8.8%) were singletons, only represented by one sequence in one sample.
Singleton clusters were relatively uniformly distributed, accounting for ~50% of the

clusters in each phylum. Two exceptions were within candidate division ODl and

89

0.01 0.05 0.12

(Specie3) (GenuS) (Family)
Reads Uncl OTU Chaol H’ OTU Chaol H’ OTU Chaol H’
BC 4442 1715 1456 3221 5.80 814 1362 4.92 445 620 4.34
EHS 4153 232 1628 5183 6.30 924 2018 5.33 478 716 4.50
JF 4796 1750 1688 3616 6.44 896 1451 5.56 459 602 4.78
WM 4127 1069 2245 6178 7.24 1372 2567 6.44 710 1088 5.59
CB 3508 909 1648 4205 6.82 978 1752 6.02 519 750 5.33
GMl 4202 207 1574 4061 6.46 810 1512 5.39 417 615 4.69
GM2 4287 260 1919 5856 6.85 1079 2118 5.94 556 903 5.18
FL9 5838 1561 2793 6734 7.40 1597 2574 6.60 761 1039 5.66
FL10 4245 141 1463 4143 6.21 734 1657 5.07 365 577 4.24
1?er 4583 1285 2825 8543 7.61 1820 3568 6.95 938 1479 6.05

 

 

 

 

 

 

Table 4.1. Total sequence reads, number of unclassified (Uncl), corresponding
number of clusters (OTUS), Chaol, and Shannon Diversity GI’) values for three
cluster cutoff dissimilarity values (0.01, 0.05, and 0.12) among sediment samples
from nine sites.
Chlamydiae where singletons constituted ~67% of the clusters. Singleton distribution
among sites was not uniform with F L1 1 expressing 17.5% singletons, while in FL10 they
only comprised 5.3% of the total population. Of the 8031 clusters, 2375 were represented
by tags in at least two samples. Surprisingly, there were no “cosmopolitan” clusters that
were found in all ten samples. One dominant cluster accounted for 1068 sequences, of
which 1063 occurred in Barrow Canyon, accounting for 24% of the sequences identified
in this sample. These sequences exhibited S_ab scores of 0.26-0.29 using the RDP
SEQMATCH tool to the closest known isolates and closest BLAST matches of 79-86%%
to the public database of all environmental sequences.

Two independent pyrosequencing runs were performed on Cascadia Basin
sediments. Chaol values for the two runs at 0.05 dissimilarity were 1752 and 1779. The

duplicate runs claded tightly together when both were included in a Chao-Sorenson

neighbor joining tree. A total of 439 clusters were shared between runs (72% of all

90

 

1800

 

IBC

1600 e FL11

 

 

 

 

1400

 

 

1200

 

1 000

800

 

 

 

600

400

 

 

200 -—- —~

 

 

 

0'1

’9 0
24°

 

Figure 4.2. Comparison of relative abundances based on the RDP Classifier between
the two most contrasting sites according to estimated richness and diversity, Florida
Bay 11 and Barrow Canyon.

91

   
 
 
 
  
  
    

 

 

30001 .
Species -P Florida Bay 11 .
. -- Washington Margin
+ Florida Bay 9
-- Cascadia Basin
25 0
O l -0- Gulf of Mexico 2
-0- East Hanna Shoal
-o- Juan de Fuca
2000, Genus ‘0' GUlf Of Mexico 1
-o- Barrow Canyon
\ ' — Florida Bay 10
1500' ‘
\ Family
I
I
1000* I
I
I
I
5001 : :
I I
I l | 4.0 ‘34“ r S
I l I
_ l . _ _ l . . v . j . '4 t - . . . . _ . .
0.01 0.05 0.12 0.20

Figure 4.3. Phylogenetic architecture based on a normalized sample size of 3,500
randomly selected sequences per site. Dissimilarity values for species, genus, and
family are noted on the x-axis while number of sequences are represented on the y-
axis.

sequences) while 907 clusters were represented by only one sequence. The abundance of
the shared clusters exhibited a standard deviation of 1.95 sequences between replicates.
Diversity Characterization

The majority of sequences belonged to the F irmicutes, 5, y, and a-Proteobacteria,
which constituted 70% of the total classified sequences (Figure 4.4). The Acidobacteria,

Actinobacteria, and Verrucomicrobia were also relatively abundant at 15% of the sum.

92

According to the RDP Classifier 9181 sequences (20.8%) out of 44,181 sequences were
unclassified.

Isolate best hit matches were analyzed using the RDP SEQMATCH tool for all
sequences. The F irmicutes exhibited the highest dissimilarity from current isolates but
were also represented by the highest isolate diversity, represented by the length of the x-
axis (Figure 4.5) with phylum calls based on RDP classification of the best hit match.
The 'y-and B-Proteobacteria were the most extensively cultivated (12.5%) while the
Planctomycetes, Spirochaetes, Gemmatinomadetes, Chloroflexi, and C yanobacteria were
among the most poorly cultivated.

The frequency distribution (Figure 4.6) illustrates the large y-Proteobacteria crest
at S_ab of 0.26-0.29 which corresponds to the single cluster of 1063 sequences found in
the Barrow Canyon sample. This distribution also illustrates a bi-modal frequency, with
an increase at S_ab of 0.5-0.7 (pairwise identity of 87-94%), illustrating a gap in isolate
coverage. Three distinct patterns of isolate coverage were identified with high frequency
exact matches and unirnodal distribution in Florida Bay 10 versus Barrow Canyon which
exhibited a high abundance of distant matches. A bi-modal distribution is also illustrated
by the sequences from Washiington Margin, with poor coverage in the mid-range, at
species-level diversity.

Cluster representative sequences were subjected to BLAST analysis according to
the abundance distribution (Figure 4.7). Singletons averaged 92.5% (sd=3.6%) BLAST
nearest match identity, clusters that contain <0.1% of sequences at each site without
singletons averaged a 95.1% (sd=3.2%) identity while those that contain >0.5% of

sequences at each site averaged a 98% (sd=1.5%) identity. A total of 1456 sequences

93

were highly divergent (<85% BLAST identity) from any environmental sequence.
Average singleton identity ranged between Juan de Fuca (91.8%) and Florida Bay 10
(94.8%).

Chao-Sorensen estimated abundances were used for the construction of a
UPGMA tree rooted in the community of a Ghana agricultural soil (data not shown) to
determine the relationship among sites (Figure 4.8a). No overall spatial or depth
component was apparent among sites. The two depths from the same Gulf of Mexico
sample did not tree together when singletons were included. East Hanna Shoal, an Arctic
sediment, grouped together with the shallow, warm Florida Bay 9 site. To determine the
influence of singletons on treeing, these rare sequences were removed to construct
another tree (Figure 4.8b). Three topological changes occur: Gulf of Mexico 1 and 2
group together, Cascadia Basin joins with Barrow Canyon, and Washington Margin and
Juan de Fuca diverge.

In order to test whether pyrosequencing could capture anaerobic ammonium
oxidizing bacteria, unclassified sequences from the RDP 10.0 release classifier were
selected from each site. BLAST was performed and sequences with >90% to known
anammox bacteria were selected. Sixteen anammox sequences were identified in six
samples and were closely related to the Ca. Scalindua type anammox, most notably Ca.

Scalindua brodae (Figure 4.9).

94

 

 

 

BC JF WC GMl GM2 FL9 FLll

CB

I Fibrobacteres
Cyanobacteria

i Chlamydiae

a Chloroflexi

I Spirochaetes

I Nitrospira

I Deferribacteres

I Planctomycetes

I Bacteroidetes

I Verrucomicrobia

I Acidobacteria

I Epsilon

I Actinobacteria

I Gamma

I Delta

I Beta

I Alpha

I Firmicutes

I Unclassified

Figure 4.4. Relative abundances illustrating the phylogenetic composition of the
sequences obtained at each site based on the RDP Classifier at 50% bootstrap

confidence threshold.

95

J

% 4% 6% 8%

 

.._Mfiia/ _ 7

 

 

90999009.“;

 

 

 

 

       

9.9030990 005890.15 J .. .. 1.5.48.
«9698.30. _ 5.988.... ..an11....
0900.90 0:90 .5. ...... . - . ...- 1.9.50.5...24 .. ..

W: “Quiz . . :11 M............ “._.-.....z.“ ... .v.............. . ...va . “ .... . .. . 1..-. .
1. n1..1U..1H1.?££H .. -.....rll
. ...fln1\hw.u.... . . .
9.9030990 09900.13... . 4...... ..._.....111 .... ...
_....RA».....: ... .... .15.... .. .
7.35.1.1... . .Jm.1um....u._..._r._........Fm.........:. . 1..
9.9000003“. . ...... 31.....111111fi151uuni» ._..... .. .....1 . .- .
“ognzhlfxlr‘fi .I..‘\ u..\ .. nauuluh‘uo.m\u l.1.Al|-—b whom—I. .... \~I.....II. .l. ....
firm... . ........x........~..318.1nlhai. Wynn. dawn: . . .
_lnr‘z ..Prn... “ma...” .. x... ... u. “...... . ...”...Iol...
832:3. ..._.....a... .. .1133... 1.1.5.1.... .............
... 5. ....
00.90000.0_u___.m.m..1..4 ..... .... . . . .. .. .
....3 .
9.9030990 c0__00w_ ._...” .....I............
9.9000090... 9.00.1... 1. 11...... .............. ... 1
....Wuflil... ...I . ...... f I‘M-.02... . u.
0.9.900 _.0too ..:...m. .30....11 ....u .......... I... x. .1... . -

 

9.90000..0>0. 090 E920 _m.
9.9030990 98 _ _.. E...

09099900m =5...» .....me1... ..

M. a... .4..........

9.9030990 0:03 .51....5...5..I.M... ..
_Hm .....1...u.nr...u.1n.... ..

9.9020994 _....
9.9000093. _0

on"

1

... .S‘
......

I ”O "

00‘“.- :‘ouoq’w to. as I...

    

090‘-

   

     

 

:5 o .
04'1000 Io. III-cubcluoobto 0000.0»..00 0

on I. I D
II

......8.... .... Fig.9”... ...WNMHQJ:

0.

6592 2.5.62 .80 .o 980 81m .9.

0.2

Percentage

lates with S ab
based frequency distribution highlighted on the right. Dashed line indicates an

iso

Figure 4.5. NEO plot of RDP best bit matches from all samples to

late.

score by phylum. One column represents a best match to one iso

S_ab score.

dentity based on the

II'WISC l

estimated 85% pa

96

 

 

 

 

 

 

 

 

_--- Barrow Canyon 8
— Florida Bay 10 " 800
_ Washington Margin "
600
1068 sequences
79—86% pairwise identity to + 5‘
closest environmental sequence 5
400 3
9
LL.
l 200
i
'1
I “g
0
0.2

 

RDP S_ab Score to Closest Isolate

Figure 4.6. Frequency distribution of S_ab scores illustrating differences in isolate
coverage among three sites. Dashed line indicates an estimated 85% pairwise
identity based on S_ab scores. The 1068 sequences that comprise the peak at low
RDP S_ab score consist of 1063 sequences exclusively from Barrow Canyon.

97

100

(O
01

(O
O

BLAST SCORE
00
01

80

75

70

 

'

 

 

 

 

 

_ ‘”“““‘ _I
“'2'.“ v.2,— :1
-Wm --

4W“. “ ‘M 2:“

 

own-mama“... C

._4...

C13“???- nu nxflfll'lafia
.‘ I! M 3- I'd" “VJ-19.
C nam Aim-53s! an.
" ' "8334- Mil-"5“;
- -~‘~1.z.v.s<. w-.~u..-.: . '1
C .mmw .mnx;
‘- 1: . an»; “same rt.»

C. “.11 sud- .mm tam

..‘t-«V'u‘; 21‘ .2185 “I ’

‘-~ Cum. seamen-41' n:-

.v “.26 -."-;‘:€3G a

“MM-Lia, 0 row .

‘- ‘b 0‘.» Sin-tan)”. {113.9

0 Acidobacteria

0 Actinobacteria
0 Aquificae

0 Bacteroidetes

.~ 0 Chlamydiae

 

O 0-- ...-.00.--. .- .mmmma‘. m
'W~ - -. ...— -——-o-—-—-¢.dv—'H-0— _—
D O 0.. om 0.. o. a. to #- No. on...

.00 I. u- -o 00. on
O O .6“ O 0 ¢ .0 0 O
O O o o o o
00 o
0 0 e
O O

 

 

 

e Chloroflexi

- Defieribacteres
o Fibrobacteres
0 F irmicutes

0 Fusobacteria

' Nitrospira

' Planctomycetes
0 Alpha

0 Beta

' Delta

' Epsilon

0 Gamma

- Spirochaetes

«- Verruco

Figure 4.7. BLAST best match scores for each cluster representative sequence based
on the cluster representative sequence RDP classifier results.

98

8°°'" Gulf1
300'“ Gulf2

 

 

 

 

 

 

 

 

 

180 m Barrow Canyon
m Juan de Fuca

 

 

dflim— Wash. Margin
—‘§9—'“— East Hanna Shoal

 

 

 

 

 

 

 

5 m Florida Bay 9
5 m Florida Bay 10

 

 

 

 

——~'fl"— Florida Bay 11
3089 m Cascadia Basin

 

 

Ghana Soil

 

 

 

Wash. Margin

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Juan _de Fuca

—— Gulf 1
—— Gulf 2

Cascadia Basin
Barrow Canyon
East Hanna Shoal

— Florida Bay 9
—— Florida Bay 10

 

Florida Bay 11

 

Ghana Soil

Figure 4.8. Chao-Sorensen cluster abundance based UPGMA tree at 5%
dissimilarity rooted with an agricultural soil from Ghana with (A) all clusters and
(B) without singletons. Depth below water surface is noted.

99

 

WC DRB9O
F L10 ANIZ3
WC C7GWP

EHS CYSAK

 

 

 

 

 

WC El 81F
WC DASll

WC DF76A
— WC BPLF R
Scalindua brodae

WC DRB9O
CB ETPCD

CB AMWFW
GM2 CEJCV

“ IF CCNXO

wc DTQBM
‘ I l we CLSSO
WC CKDF3

Scalindua wagneri

   
  
   
  

 

 

 

 

 

 

 

 

Brocadia fulgida

Brocadia anammoxidans

 

 

 

 

 

 

 

Kuenenia stuttgartiensis

 

 

 

Jettenia asiatica

'——' Anammoxoglobus
propionicus

 

 

 

Figure 4.9. Relationship of identified sequences with >98% sequence homology to
the known anammox bacteria. Site designations are referred at the beginning of
each sequence name.

100

DISCUSSION
Community Diversity

Pyrosequencing targeting the 16S rRNA gene for community analysis is evolving
into a powerful tool providing an unparalleled view into the diversity of the microbial
world. The high coverage V4 hypervariable region primers used in this study are
supported by in-silico UNIFRAC analysis (Liu et al., 2007), result in amplicons that have
low insertion and deletion frequencies for simplified alignment, and have high resolution
and accuracy for deep taxa classification using the RDP Classifier (Wang et al., 2007).

Our results indicate that SSU rRNA pyrosequencing can be used to assess and test
hypotheses concerning microbial diversity and relative abundance among different
environments. Specifically, our data indicates that 7.9% of the 44,000 sequences
examined were identical to those recovered from cultivated bacteria, challenging the
notion that “less than 1% of bacteria are cultivated”. Among our sequences, 3.3% of were
highly divergent with equal to or less than 85% BLAST identity against SSU rRNA
sequences in the public database. These represent highly divergent taxa and are not
uniquely rare, as the case with the large cluster in our Barrow Canyon sediment. As such,
pyrosequencing appears to be a powerful method for identifying poorly characterized
taxa that may play a large metabolic role in the environment.

Singletons, identified as sequences within a cluster at 5% dissimilarity that are
present only in one sample, represented 8.8% of all sequences obtained in the study,
compared to 18.1% found in a deep clone library from a soil sample (Elshahed et al.,
2008). Both Chlamydiae and candidate division 0D] exhibited the highest proportion of

singletons (~67%) in our study, in comparison to the highest proportion in

101

Planctomycetes in the soil biosphere (Elshahed et al., 2008). The increasing sequence
divergence from the database as rarity increases is also in agreement with the clone

library based study and a marine sediment pyrosequencing study (Sogin et al., 2006).

The Rare Biosphere

When singletons are removed the topological changes in the cluster based tree
illustrate differences between the overall diversity and the dominant “functional capacity”
of the sediment. As rare members, singletons are not expected to contribute a large
fraction of the total sediment metabolism. Several hypotheses have been presented to
explain the presence of these rare members. They may represent a “genetic reservoir”, a
temporally-based population that can respond to changes in the environment and
potentially become dominant when conditions are appropriate (Sogin et al., 2006).
Secondly, they may be transients, in our case carried by currents, which are not currently
adapted to the site conditions. Third, they may be less adapted “fuzzy” species,
evolutionary remnants that are currently out-competed in all global ecosystems but are
able to survive (Elshahed et al., 2008). In the first two cases we would expect that
comparison to the public database would lead to identities similar to the distribution of
the less rare members. Exceptions to this would be environments that are not well
represented by the current database, such as the deep biosphere.

Our data shows that the rare biosphere is indeed a divergent population that is
under-sampled in all environments. This does not appear to support the rare members as
transient species or a genetic reservoir defined as a population that responds to changing

environmental conditions. If the rare biosphere was dominated by transients we would

102

expect to find less unique sequences that are present only in one site and BLAST
identities would be less divergent, as the database reflects a large span of biogeochemical
conditions. It appears most likely that these are either bacteria suited to a specific niche
that has not been well sampled, or that they are “fuzzy” species that may play a minor
role in sediment nutrient cycling. However, it is impossible to discard the notion that
large scale environmental changes may result in proliferation of the rare biosphere (Sogin
et al., 2006), thus supporting the notion of the “genetic reservoir”. For example, warming
global temperatures may provide an environment that favors certain enzyme systems over
others, which may currently reside in the rare biosphere. Conversely, nutrient loading
may favor rare members with higher enzyme Km values that are currently out-competed
for limiting resources.

While the average identities were lower among the singletons, there was a notable
exception in Barrow Canyon. The most dominant overall cluster consisted of 1063
sequences with a low BLAST nearest neighbor identity of 79-86%. This highly
evolutionary distinct lineage represents 24% of all sequences in BC and, due to its
prevalence, may play a significant role in the microbial loop. Likely examples of
microbial endemism, these dominant members are also identified in low abundance in

two other sites.

Selective Effects in Florida Bay
Sampling of the two Florida Bay sites allowed us to investigate the potential
causes behind different relative abundances in certain dominant phyla. Due to the

influence of Gulf of Mexico waters, the otherwise oligotrophic Florida Bay is subjected

103

to a phosphorus gradient that is reflected in surface water and seagrass nutrient levels.
Florida Bay 11 is most influenced by Gulf waters with higher phosphorus concentrations
and lower N:P ratios. Conversely, Florida Bay 9 is subjected to onshore-offshore
dynamics with lower P availability, higher dissolved organic carbon (DOC), and
markedly lower salinity (Boyer et al., 1997; Fourqurean and Zieman, 2002). Our cluster
trees, with or without singletons, show that the Florida Bay 9 and 11 communities are
divergent, with F L10 most closely associated with PM 1. This is in contrast with the site
grouping based only on unfilitered dissolved organic matter (UDOM) characteristics in
the water column (Maie et al., 2005). We expected that F L10 would be an intermediate
site in terms of its impact from the Gulf of Mexico. However, FL10 has exhibited P
concentrations a third less than FL9 in another study (Chambers and Pederson, 2006)
which likely contributes to much lower Chaol and Shannon Diversity values. Since this
parameter falls out of the previously documented east-west trends (DOC, CzN, N:P,
Salinity, etc.) we will restrict further comparisons to either end of the spectrum, F L9 and
PM 1.

Overall, both Florida Bay sites have the highest Chaol and Shannon Diversity
estimates of the ten sediments, suggesting that Florida Bay sediments are more
taxonomically rich and diverse than many aquatic sediments, a supposition supported by
an earlier study (Hewson et al., 2002). Compared to FL9, the larger FLll Chaol and
Shannon Diversity values may be partially influenced by a higher viral abundance, as
virus particles are typically higher in, and limited by P (Hewson et al., 2001; Hewson and
Fuhrman, 2002). Nutrient amended mesocosms in Florida Bay have indeed shown that

higher bacterial diversity and richness was accompanied by increased viral abundance

104

(Hewson et al., 2002), likely due to predation based on encounter probability (Thingstad,
2000). Another selective influence is the abundance of grazers, such as Bdellovibrio,
which comprise 0.5% of the total community at FLll with only one sequence at FL9.
Therefore, increased diversity driven by predation appears to be consistent with our study
where apparent viral driven and grazing selection is not overshadowed by increased P-
limited microbial speciation due to constrained horizontal gene transfer (Souza et al.,
2008)

The outcomes of both phosphorus and predation selective effects are illustrated in
the differences in the relative abundances between sites. If predation is playing a role in
shaping the microbial community structure, then we expect that changes in the relative
abundances of certain bacteria would follow trends observed in previous studies. As
expected, the B-Proteobacteria were equal in number between F L9 and PM 1, as they
have been shown to be opportunistic resource competitors able to compete in a variety of
nutrient conditions with little grazing pressure. In contrast, Actinobacteria were three
times more abundant in F L1 1, comprising 4% of the population. While Actinobacteria
are generally not directly affected by nutrient status, they do profit from the removal of
competitors through grazing pressure driven by higher P availability (Salcher et al.,
2007). Bacteroidetes, who have been shown to profit from a high nutrient load and are
triggered by grazing pressure (Salcher et al., 2007), composed 3.5% of the community at
FL] 1, twice that of F L9. Of course, site specific nutrient conditions, other than
phosphorus availability, will influence microbial community composition. One example
is the overall two-fold higher relative abundance of the putative sulfur cycling 6-

Proteobacteria at FL9, comprising 20% of all sequences at this site. Overall, the 5-

105

Proteobacteria are the most dominant bacteria in both sites and represent 28% and 20%
of all sequences in FL9 and FL] 1, respectively. Overall, these correlations with earlier
studies suggest that predator pressure, not uniquely phosphorus availability, is a
controlling force in these shallow Florida Bay sediments and may play a role in shaping
marine microbial community structure on a larger scale.

Overall there were no apparent large scale spatial or depth related trends among
all sites. The grouping of the two depths of the same Gulf of Mexico core after the
singletons were removed illustrates one of the issues that pyrosequencing presents.
Namely, what sequencing depth do we choose in order to illustrate the grouping of sites?
If an approximation of the true diversity is the goal, then the inclusion of singletons
appears to be appropriate. However, if investigating the potential functional or
biogeochemical relationship likely reflected by the dominant members, singletons may be
excluded. These issues can be addressed in the future by comparing metatranscriptome

and 168 targeted pyrosequencing based data.

ACKNOWLEDGEMENTS
Appreciation goes to Woo Jun Sul who developed the pyrosequencing primers,
protocol, performed the PCR runs and was integral in workflow development. The staff
of the RDP, particularly Dr. Jim Cole and Qiong Wang, were responsible for developing
the bioinformatic tools and pyrosequencing pipeline used in this study. Dieter Tourlousse
also provided helpful insights into working with the data. Dr. Allan Devol was
responsible for collecting the Arctic marine samples. Dr. Sonia Tiquia provided the Gulf

of Mexico samples and Dr. Joe Boyer collected the Florida Bay samples.

106

REFERENCES

Acinas S.G., Klepac-Ceraj V., Hunt D.E., Pharlno C., Ceraj I., Distel D.L., and M.F.
Polz. 2004. Fine-scale phylogenetic architecture of a complex bacterial community.
Nature 430:551-554.

Ashby M.N., Rine J., Mongodin E., Nelson K.E., and D. Dmster-Denk. 2007. Serial
analysis of rRNA genes and the unexpected dominance of rare members of microbial
communities. Appl. Environ. Microbiol. 73:4532-4542.

Baas Becking L.G.M. 1934. Geobiologie of inleiding tot de milieukunde. The Hague,
the Netherlands: W.P. Van Stockum & Zoon (in Dutch).

Boyer J.N., Fourqurean J.W., and R.D. Jones. 1997. Spatial characterization of water
quality in Florida Bay and Whitewater Bay by multivariate analysis: zones of similar
influence. Estuaries 20:743-758.

Campbell B.J., Engel A.S., Porter M.L., and K. Takai. 2006. The versatile epsilon
proteobacteria: key players in sulphidic habitats. Nat. Rev. Microbiol. 4:458.

Cannone J.J., Subramanian S., Schnare M.N., Collett J.R., D’Souza L.M., Du Y.,
Feng B., Lin N., Madabusi L.V. Muller K.M., Pande N., Shang 2., Yu N., and RR.
Gutell. 2002. The Comparative RNA Web (CRW) Site: an online database of
comparative sequence and structure information forribosomal, intron, and other RNAs.
BMC Bioinformatics 3:2.

Chao A., Chazdon R.L., Colwell R.K., and T.J. Shen. 2006. Abundance-based
similarity indices and their estimation when there are unseen species in samples.
Biometrics 62:361-371.

Chambers R.M., and K.A. Pederson. 2006. Variation in soil phosphorus, sulfur, and
iron pools among south Florida wetlands. Hydrobiologia 569:63-70.

Curtis T.P., Sloan W.T., and J.W. Scannell. 2002. Estimating prokaryotic diversity and
its limits. Proc. Natl. Acad. Sci. USA 99: 10494—10499.

Elshahed M.S., Youssef N.H., Spain A.M., Sheik C., Najar F.Z., Sukharnikov L.O.,
Roe B.A., Davis J.P., Schloss P.D., Bailey V.L., and L.R. Krumholz. 2008. Novelty

and uniqueness patterns of rare members of the soil biosphere. Appl. Env. Microbiol.
74:5422-5428.

Fourqurean J.W., and J.C. Zieman. 2002. Nutrient content of the seagrass Thalassia
testudinurn reveals regional patterns of relative availability of nitrogen and phosphorus in
the Florida Keys, USA. Biogeochem. 61:229-245.

107

Hagstriim A, Pommier T., Rohwer F ., Simu K., Stolte W., Svensson D., and U L.
Zweifel. 2002. Use of 16S ribosomal DNA for delineation of marine bacterioplankton
species. Appl. Environ. Microbiol. 68:3628—3633

Hewson 1., O’Neil J.M., Fuhrman J.A., and W.C. Dennison. 2001. Virus-like particle
distribution and abundance in sediments and overlying waters along eutrophication
gradients in two subtropical estuaries. Limnol. Oceanogr. 47:1734-1746.

Hewson 1., and J.A. Fuhrman. 2002. Virioibenthos abundance and production in neritic
and basin sediments of Southern California. Eos 83:0831C-29.

Hewson 1., Vargo G.A., and J.A. Fuhrman. 2003. Bacterial diversity in shallow
oligotrophic marine benthos and overlying waters: Effects of virus infection,
containment, and nutrient enrichment. Microb. Ecol. 46:322-336.

Huber J.A., Welch D.B.M., Morrison H.G., Huse S.M., Neal P.R., Butterfield D.A.,
and M.L. Sogin. 2007. Microbial population structures in the deep marine biosphere.
Science 318: 97-100.

Liu Z., Lozupone C., Hamady M., Bushman F.D., and R. Knight. 2007. Short
pyrosequencing reads suffice for accurate microbial community analysis. Nucleic Acids
Res. 35:120.

Lozupone C.A. and R. Knight. 2007. Global patterns in bacterial diversity. 104:11436-
11440. Proc. Natl. Acad. Sci.

Maie N., Yang C., Miyoshi T ., Parish K., and R. Jaffé. 2005. Chemical characteristics
of dissolved organic matter in an oligotrophic subtropical wetland/estuarine ecosystem.
Limnol. Oceanogr. 50:23-35.

Pedros-Alio C. 2006. Marine microbial diversity: can it be determined. Trends.
Microbiol. 14:257-263.

Pernthaler, J. 2005. Predation on prokaryotes in the water column and its ecological
implications. Nat. Rev. Microbiol. 3:537—546

Pommier, T., Pinhassi J., and A. Hagstroem. 2005. Biogeographic analysis of
ribosomal RNA clusters from marine bacterioplankton. Aquat.Microb. Ecol. 41:79—89

108

Rusch D.B., Halpern A.L., Sutton G., Heidelberg K.B., Williamson S., Yooseph S.,
Wu D., Eisen J.A., Hoffman J.M., Remington K., Beeson K., Tran B., Smith H.,
Baden-Tillson H., Stewart C., Thorpe J., Freeman J., Andrews-Pfannkoch C.,
Venter J.E., Li K.E., Kravitz S., Heidelberg J.F., Utterback T., Rogers Y.-H., F alcon
L.I., Souza V., Bonilla-Rosso G., Eguiarte L.E., Karl D.M., Sathyendranath S., Platt
T., Bermingham E., Gallardo V., Tamayo-Castillo G., Ferrari M.R., Strausberg
R.L., Nealson K., Friedman R., Frazier M., and J.C. Venter. 2007. The Sorcerer 11
Global Ocean Sampling Expedition: Northwest Atlantic through Eastern Tropical Pacific.
PLoS Biol 5:e77.

Salcher M.M., Hofer J., Hornak K., Jezbera J., Sonntag B., Jaroslavv, Simek K.,
and T. Posch. 2007. Modulation of microbial predator-prey dynamics by phosphorus

availability: Growth patterns and survival strategies of bacteria phylogenetic clades.
FEMS Microb. Ecol. 60:40-50.

Schmid, M.C., Risgaard-Petersen N., van de Vossenberg J., Kuypers M.M.M., Lavik
G., Petersen J., Hulth S., Thamdrup B., Canfield D., Dalsgaard T., Rysgaard S.,
Sejr M.K., Strous M., Op den Camp H.J.M., and M.S.M. Jetten. 2007. Anaerobic
ammonium-oxidizing bacteria in marine environments: widespread occurrence but low
diversity. Environ. Microbiol. 9: 1476-1484.

Sogin M.L., Morrison H.G., Huber J.A., Welch D.M.," Huse S.M., Neal P.R., Arrieta
J.A., and G.H. Herndl. 2006. Microbial diversity in the deep sea and the underexplored
“rare biosphere.” Proc. Natl. Acad. Sci. 103:12115-12120.

Souza V., Eguiarte L.E., Siefert J., and J.J. Elser. 2008. Microbial endemism: does
phosphorus limitation enhance speciation? Nat. Rev. Microbiol. 6:559-564.

Thamdrup B., and T. Dalsgaard. 2002. Production of N2 through anaerobic

ammonium oxidation coupled to nitrate reduction in marine sediments. Appl. Environ.
Microbiol. 68:1312-1318.

Thingstad T.F. 2000. Elements of a theory for the mechanisms controlling abundance,
diversity, and biogeochemical role of lytic viruses in aquatic systems. Limnol. Oceanogr.
45:1320—1328

Wang Q., Garrity G.M., Tiedje J.M., and JR. Cole.. 2007. Naive Bayesian classifier
for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ

Microbiol 73:5261-5267.

Whitfield J. 2005. Biogeography: is everything everywhere? Science 310:960-961.

109

CHAPTER V

QUANTITATIVE PCR OF FRESHWATER ANAMMOX IN
HAWAIIAN TARO FIELDS

ABSTRACT

The identification of anammox bacteria in a variety of freshwater habitats has
suggested that anammox may be responsible for nitrogen removal in these systems.
However, there is currently little evidence concerning the abundance of these microbes in
flooded agricultural sediments. A previous study indicated that up to 80% of added
fertilizer nitrogen could not be accounted for by volatilization, plant uptake, or
denitrification, indicating that anammox activity may be responsible for the N loss.
Sediments from a variety of Hawaiian taro fields were analyzed for anammox bacterial
abundance using two primer sets targeting the “marine” genera Scalindua and the
“freshwater” Brocadia, Kuenenia, and Jettenia. Total estimated anammox numbers were
one to two orders of magnitude than those found in deep marine sediments. Among the
agricultural plots, the newly flooded taro field following fertilization exhibited the
highest estimated counts. Potential N removal rates calculated from the range of
published cell-specific activities showed that anammox may be responsible for the
removal of 2 to 763 kg’1 N ha'1 yr". An estimated rate in Hanalei river sediment showed a
much higher N removal potential, suggesting that this system may harbor a very active

anammox population.

110

INTRODUCTION

The importance of anaerobic ammonium oxidation is becoming more well-
defined in marine systems, with the process responsible for between 0 and 80 percent of
total N2 production in a variety of basins (Thamdrup and Dalsgaard, 2002; Risgaard-
Petersen et al., 2004; Engstrom et al., 2005). However, the significance of this process in
freshwater sediments including flooded agricultural soils has been largely unexplained.
Only a few studies have referenced anammox bacteria or the process in these
environments (Jetten et al., 2003; Trimmer et al., 2003; Schubert et al., 2006; Zhang et
al., 2007). In a freshwater lake, anammox was shown to contribute up to 13% of total N2
production in Lake Tanganyika (Schubert et al., 2006), but no literature describing
anammox in flooded agricultural systems is currently available.

The contribution of anammox to total N loss in flooded agricultural systems is of
great importance in determining the appropriate fertilization application amounts and
methods such that N losses are mitigated. This is important not only in terms of the
financial burden but also the eutrophication of surrounding waters resulting from excess
fertilizer inputs. Ammonium is relatively abundant due to organic matter mineralization
and fertilization, there is a thin oxic soil layer at the water/soil interface, and the subsoil
remains anoxic, conditions that would appear suitable for anammox bacteria.

Anammox activity has been confirmed in rice paddies, but the overall
contribution to total N2 production (anammox + denitrification) is thought to be low
(Suwa, ISME, 2008). Growing season length and treatment between cropping seasons has
a direct impact on the stability of the anoxic environment. Rice has a growing season of 3

to 7 months with the soil becoming aerated in between crops. As a result, the anammox

111

anoxic environment is transient, which may affect not only the anammox density but the
metabolic resilience of the bacteria as they are stressed. Taro is a tropical plant and a
staple food in Polynesian cultures whose environment should be well suited for anammox
bacteria with along crop cycle (14-16 months) and continuous soil flooding.

In an experiment testing N availability from flooded taro fertilized with fish-bone
meal on Oahu, Hawaii in 2003-2004, a rapid release of NH4+ was observed from the
organic amendment applied to the surface soil (0-15 cm depth) immediately before
flooding. More that 80% of the mineralized NH4+ could not be accounted for as taro plant
uptake accounted for only 10% of the added N and floodwater nitrate and ammonium
concentrations remained low, despite high soil NH4+ concentrations. This suggests that
volatilization is low, which was supported by earlier studies (Cao et al., 1984; Fillery et
al., 1984). Additionally, losses to denitrification appear to be low due to soil anaerobiosis
which limits nitrate and nitrite production. Experiments have shown that denitrification
losses are low when fertilizer was applied to already flooded soils (Mosier et al., 1998).
Unlike rice systems where rhizosphere nitrification occurs in the bulk anoxic soil due to
fine, diffuse roots (Arth et al., 1998), the tuber root systems of taro plants are less likely
to support significant surface area for nitrification.

The determination of anammox activity in these flooded agricultural systems is of
interest due to its impact on N dynamics, crop productivity, and modeling N losses. In
this study seven flooded systems are tested for the presence of anammox bacteria that
may be responsible for these observed N losses. Two separate quantitative PCR assays

targeting the marine Scalindua sp. and the freshwater Kuenenia stuttgartiensis, Brocadia

112

anammoxidans, and Brocadiafulgida related anammox are used to assess their density in

these soils.

METHODS
Quantitative PCR Assays

Quantitative PCR targeting the 16S rRNA gene of the freshwater anammox
Kuenenia stuttgartiensis, Brocadia anammoxidans, and Brocadia fulgida along with the
marine anammox Scalindua sp. was performed using the SYBR Green assay (Applied
Biosystems). The primer sequences were as follows: 818F (5’-ATG GGG CAC TMR
GTA GAG GGG TTT- 3’) and 1066R (5’-AAC GTC TCA CGA CAC GAG CTG-3’)
(Tsushima et al., 2007). Each 25 uL reaction mixture contained 12.5 uL lx Sybr Green
Master Mix, 300 nM of each primer, 100 ug mL"l BSA (Roche), and 2.5 uL template
DNA or 101 to 108 copies of the standard plasmid carrying the target DNA of Brocadia
anammoxidans. The reaction conditions were: 2 min 50°C, 10 min 94°C, 40 cycles at 153
at 94°C and 1 min at 60°C. Standard curves were generated using a plasmid dilution
containing the 168 rRNA gene of Brocadia anammoxidans from 103 to 108 copies.

The Taqman assay described in Chapter III was used for the detection of
anammox 16S rRNA genes using primers: 541F RT (5’-GAG CAC GTA GGT GGG TTT
GTA AG-3’) and 616RRT (5’-CCT CCT ACA CTC AAG ACT YGC AG-3’). The
Taqman probe AnPrb (5’-CAG RTG TGA AAG CCT TCT GTT CAA CGG AAG-3’)
was labeled 5’ FAM and 3’ Black Hole Quencher-1® (Integrated DNA Technologies,
Coralville, Iowa). Final assay concentrations of 541FRT, 616RRT, and AnPrb were 300

nM, 900 nM, and 250 nM, respectively with l uL of genomic DNA added per well and

113

annealing and extension temperatures of 60°C. The increase in fluorescence emission
was monitored during PCR amplification using a 7700 Sequence Detector (PE Applied
Biosystems). The CT values obtained for each sample in both assays were compared to

the standard curve to determine the initial 16S rRNA gene copy number g'1 wet sediment.

Sample Description and DNA Extraction

Soil samples were collected from seven sites in five different taro patches;
flooded taro (Oahu), dry taro (Oahu), newly flooded taro (Oahu), and anaerobic taro
where anaerobic conditions have been maintained for the longest period (Oahu), flooded
taro (Hanalei) along with a sediment sample from the Hanalei River (Kauai), and a duck
pond adjacent to the Hanalei River (Kauai).

Genomic DNA from 0.5 g sediment was isolated using the MoBio Ultraclean soil
DNA kit (MoBio Laboratories, Inc.) and further purified by electrophoresis in 0.8%
Cambrex SeaPlaque low melting temperature agarose at 4°C. Genomic DNA bands were
gel-extracted, treated with gelase (Epicentre) per manufacturer’s instructions, and
concentrated using Microcon columns. DNA concentrations were determined by

absorbance at 260 nm.

RESULTS AND DISCUSSION
Anammox bacteria related to Kuenenia stuttgartiensis, Brocadia anammoxidans,
Brocadia fulgida, and Jettenia asiatica were enumerated using the Sybr Green assay
developed by Tsushima et a1. (2007). Melting curve analysis showed that there was only

one observable peak, no other detectable peaks that were associated with primer-dimer

114

artifacts or other non-specific PCR amplification were observed. A standard curve was
constructed using a Brocadia anammoxidans 16S rRNA gene containing plasmid dilution
series (Figure 5.1). The resulting slope (-3.03) approached the value of perfect
amplification efficiency (-3.3). The strong linear relationship (R2=0.98) between
threshold cycle number and target copy number demonstrated the consistency of the
assay. Error increased considerably at 103 copies, indicating approaching detection limit.
This assay also enumerates the marine anammox bacteria related to Candidatus
Scalindua brodae and Candidatus Scalindua wagneri with one mismatch. Therefore, the
resulting PCR efficiency of the Scalindua sp. is one to two orders of magnitude lower
(Tsushima et al., 2007).

Marine anammox related bacteria were detected in all seven samples at a density
ranging from 104 to 105 168 rRNA gene copies g'1 wet soil (Figure 5.2). The lowest
abundances were associated with the dry and newly flooded Oahu taro patches while the
highest anammox density was found in the three flooded or anaerobic Oahu taro patches.
Anammox numbers in these soils were two to three orders of magnitude lower than those
found in the deep marine sediments (mean=2* 106 copies g'1 wet sediment, Chapter III) .
Identification of these principally marine species in freshwater habitats is not surprising,
as the marine Scalindua sp. type anammox were shown to be distributed widely in the
environment (Penton et al., 2006). However, the presence of the freshwater anammox in
flooded agricultural systems cannot be overlooked as both anammox groups have been
found to co-exist in freshwater environments (Zhang et al., 2007).

Anammox bacteria counts were one to two orders of magnitude higher than in the

assay that only targeted the marine Scalindua sp. type anammox in the same samples

115

40 ' slope = -3.03
R2 = 0.98

Threshold Cycle CT
21 5 .3

#
LII

 

.—s
O

 

2 3 7: 5 6 '7 2i 6
Log l6S rRNA gene copy number g'1 wet sediment

Figure 5.1. Standard curve based on a dilution series of plasmids containing the 168
rRNA gene of Brocadia anammoxidans.

116

1x107-

1x106-

lxlOsr

16S rRNA gene copy number g'1 wet sediment

 

 

4
1x10 -
Flooded Dry Taro Newly Anaerobic Hanalei Duck Flooded
Taro (Oahu) Flooded Taro River Pond Taro
(Oahu) (Oahu) (Oahu) (Hanalei) (Hanalei)

Figure 5.2. Distribution of all anammox bacteria based on the 818F/1066R primer
set (dark grey) compared to the marine only Scalindua sp. related anammox (light
grey). Abundances are normalized to 16$ rRNA gene copy number g"1 wet sediment‘

117

(Figure 5.2). One exception occurred in the Hanalei flooded taro field sediment where
both assays detected similar numbers of bacteria. Specifically, the 818F/1066R assay
detected 1.58x105 versus 1.48x105 l6S rRNA genes g'1 wet sediment in the marine
Scalindua sp. assay. This is in contrast to the large difference between assays in the
Hanalei River sediment where the broader anammox assay identified 3.68x107 16S rRNA
genes g'1 wet sediment more anammox than the marine primer set. The range of
anammox bacteria numbers were within the range reported by Tsushima et al. (2007) in
batch culture enrichments where most assays identified 106 to 108 anammox 16S rRNA
gene copies ml".

The estimates of total anammox in each site were applied to the correlation
between the nitrogen removal rate and cell numbers reported in a batch culture
experiment (N removal rate (uM hr'l) = 3.0 x 10'6 * (anammox 16S rRNA gene copy
number ml'1)) (Tsushima et al., 2007). Predicted rates ranged from 0.47 to 110.95 nmoles
N hr'l g'1 wet sediment removed (Table 5.1).This calculation assumes that 1 g sediment is
equivalent to 1 ml of media in the batch reactor analysis. Potential rates were also
calculated based on the range of cell specific activities for anammox bacteria from
Chapter III and the literature (1-22 fmol ammonium cell day'l), in a wide range of
environments (Kuypers et al., 2003; Kuypers et al., 2005; Schubert et al., 2006; Strous et
al., 1999). These calculated values ranged from 0.01 to 77.1 nmoles N removed g'1 wet
sediment hr'l which can be directly compared with observed anammox rates (41 to 58
nmol cm'3 hr") in a 25 day old bioreactor enrichment using a riverine sediment (Zhang et
al., 2007). Overall, the estimated rates range from 0.01 to 110.95 nmols N consumed g'1

wet sediment hr'l when the total anammox 818F/1066R assay was used. Estimated

118

nitrogen loss using these estimated anammox activities ranged from 1.8 to 763.4 kg N ha'
I yr”.

The marine anammox targeted assay identified that the dry and newly flooded
taro fields exhibited the lowest anammox abundance, as would be expected due to the
current or recent aerobiosis. However, the total anammox assay disagreed, showing the
there were more anammox in the newly flooded field than the flooded taro at both the
Oahu and Hanalei sites. The reason for this disparity is not apparent. The only other
evidence showing the co-existence of the two anammox groups is from a study in Xinyl
River, China (Zhang et al., 2007). One possible explanation is that the freshwater type
anammox, who have been shown to have multiple metabolic pathways, may be able to
tolerate transient aerobic conditions better than the Scalindua sp. type anammox. Another
possibility is that an ammonium pulse resulting from the mobilization of freshly
mineralized organic matter while under aerobic conditions provides a transient
environment well-suited to anammox bacteria proliferation. The highest anammox
numbers in the Hanalei River sediment suggests a stable, anaerobic environment that is
harboring a very large anammox population almost two magnitudes larger than that
observed in the deep ocean (Chapter III). Interestingly, the flooded Hanalei taro sediment
had very similar anammox counts between the freshwater and total anammox assays,
indicating that the marine anammox accounted for 93% of the population. In contrast
they only constitute 0.4% of the community in the Hanalei River sediment. Since

amplification efficiency of the marine Scalindua sp. anammox is one to two orders of

magnitude lower in the total anammox 818F/1066R assay, it is probable that total

119

 

 

 

 

 

 

 

 

 

 

 

 

Low High Tsushima
Flooded Taro (Oahu) 0.06 1.38 1.98
Dry Taro (Oahu) 0.02 0.50 0.72
Newly Flooded (Oahu) 0.12 2.88 4.15
Anaerobic Taro (Oahu) 0.02 0.59 0.86
Hanalei River 3.08 77.05 110.95
Duck Pond (Hanalei) 0.06 1.45 2.09
Flooded Taro (Hanalei) 0.01 0.33 0.47

 

Table 5.1. Estimated nitrogen removal rates based on quantitative PCR runs using
the 818F/1066R primer set. The low and high columns show N removal rates
calculated from literature-based cell specific activities at l and 20 fmol ammonium
cell'l day". The last column is based on the N consumption rate calculation reported
by Tsushima et al. (2007). Both calculations are normalized to the reported units of

nmols nitrogen (NH4+ + NO2') consumed hr'l g'1 wet sediment”.

120

 

 

Original Values

Corrected Values

 

 

 

 

 

 

 

 

 

 

(105 copies g") (105 copies g’l)
Flooded Taro (Oahu) 6.61 7.96
Dry Taro (Oahu) 2.40 3.09
Newly Flooded (Oahu) 13.8 1.41
Anaerobic Taro (Oahu) 2.85 3.96
Hanalei River 370 3.71
Duck Pond (Hanalei) 6.98 8.31
Flooded Taro (Hanalei) 1.58 3.04

 

Table 5.2. Original and corrected estimated anammox abundances based on
quantitative PCR using the 818F/1066R primer set. Corrected abundances were
calculated by adding the difference from a two-magnitude underestimation of the
marine Scalindua sp. anammox, based on the estimated Scalindua sp. anammox
abundances using the Taqman 541F/616R primer set. All values are anammox 16S

rRNA gene copy number g’1 wet sediment”.

121

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Low* High* Tsushima* Difference
Flooded Taro (Oahu) .07 1.66 2.39 +20.4%
Dry Taro (Oahu) .03 .64 0.93 +28.3%
Newly Flooded (Oahu) .12 2.93 4.22 +1 .7%
Anaerobic Taro .03 .83 1.19 +39.0%
LOahu)
Hanalei River 3.09 77.35 1 1 1.39 +.4%
Duck Pond (Hanalei) .07 1.73 2.49 +19.1%
Flooded Taro .03 .63 0.91 +92.2%
(Hanalei)

 

Table 5.3. Estimated corrected nitrogen removal rates based on quantitative PCR
runs using the 818F/1066R primer set. Corrections were based on calculated
abundances from Table 5.2 using a two magnitude underestimation of the marine
Scalindua sp. type anammox. The low and high columns show N removal rates
calculated from literature-based cell specific activities at l and 20 fmol ammonium
cell'l day". The next to last column is based on the N consumption rate calculation
reported by Tsushima et al. (2007). Both calculations are normalized to the reported
units of nmols nitrogen (NH: + NO2') consumed hr'l g'1 wet sediment”. The last
column shows the percent increase in estimated N removal when corrected marine

Scalindua sp. counts are used.

122

 

anammox numbers are slightly higher than those reported. To correct for this, anammox
numbers indicated by the marine anammox 541F/616R assay were decreased two orders
of magnitude and the difference was added to the total anammox 818F/1066R assay
results (Table 5.2). To estimate the effect of the increase in Scalindua sp. type anammox
counts, new estimated N removal rates were also calculated (Table 5.3). The new values
reflect increases of 0.4 to 92.2% over the estimates without correction for under-reporting
the Scalindua type anammox. Sites, such as the flooded Hanalei taro, where the marine
anammox counts were close to the total anammox counts, showed the largest increase in
potential N removal rates.

Overall, the estimated rates and abundances suggest that aeration status does not
uniformly influence anammox bacteria. Rather, it may be driven more by N availability
in these environments. For example, the newly flooded Oahu taro field may benefit from
the ammonium pulse following flooding and the Duck Pond sediment may be influenced
largely by the high N load resulting from duck droppings. As such, it is expected that
anammox activities and numbers change throughout the taro growing season and that
fluctuations will occur in direct response to fertilizer application. If this is the case then
anammox may be responsible for a significant proportion of N losses in these systems

where denitrification is limited due to sub-surface fertilizer applications.

ACKNOWLEDGEMENTS
Kind thanks goes out to Dr. Jonathan Deenik of the University of Hawaii for
envisioning the role of anammox in taro soils and providing the samples and preliminary

results that led to this investigation.

123

REFERENCES

Arth 1., Frenzel P., and R. Conrad. 1998. Denitrification coupled to nitrification in the
rhizosphere of rice. Soil Biol Biochem. 30:509—5 15

Cao Z.H., Dedatta SK, and I.R.P. Fillery. 1984. Effect of placement methods on
floodwater properties and recovery of applied nitrogen (N-lS-Labeled Urea). Wetland
Rice. Soil Sci. Soc. Amer. J. 48:196-203.

Engstrt’im P., Dalsgaard T., Hulth S., and R.C. Aller. 2005. Anaerobic ammonium
oxidation by nitrite (anammox): Implications for N2 production in coastal marine
sediments. Geochim. Cosmochim. Acta. 69:2057-2065.

Fillery I.R.P., Simpson J.R., Dedatta SK. 1984. Influence of field environment and
fertilizer management on ammonia loss from flooded rice. Soil Sci.Soc. Amer. J. 48:914-
920.

Jetten M.S.M., Sliekers O., Kuypers M., Dalsgaard T., van Niftrik L., Cirpus 1., van
de Pas-Schoonen K., Lavik G., Thamdrup B., Le Paslier D., Op den Camp H.J.,
Hulth S., Nielsen L.P., Abma W., Third K., Engstriim P., Kuenen J.G., Jorgensen
B.B., Canfield D.E., Sinninghe Damsté J.S., Revsbech N.P., Fuerst J., Weissenbach
J., Wagner M., Schmidt 1., Schmid M, and M. Strous. 2003. Anaerobic ammonium
oxidation by marine and freshwater planctomycete-like bacteria. Appl Microbiol
Biotechnol 63: 107—114.

' Kuypers M.M.M., Lavik G., Woebken D., Schmid M., Fuchs B.M., Amann R.,
Jorgensen B.B., and M.S.M. Jetten. 2005. Massive nitrogen loss from the Benguela
upwelling system through anaerobic ammonium oxidation.

Kuypers, M.M.M., Sliekers A.O., Lavik G., Schmid M., Jorgensen B.B., Kuenen
J.G., Sinninghe Damsté J.S., Strous M., and M.S.M. Jetten. 2003. Anaerobic
ammomium oxidation by anammox bacteria in the Black Sea. Nature. 422:608-611.

Mosier A.R., Duxbury J.M., Freney J.R., Heinemeyer O., and K. Minami. 1998.
Assessing and mitigating N20 emissions from agricultural soils. Climat. Change
4027-3 8.

Penton C.R., Devol A.H., and J.M. Tiedje. 2006. Molecular evidence for the broad
distribution of anaerobic ammonium-oxidizing bacteria in freshwater and marine
sediments. Appl. Env. Microb. 72:1-4.

Risgaard-Petersen N., Meyer R.L., Schmid M.C., Jetten M.S.M., Enrich-Prast A.,

Rysgaard S., and N.P. Revsbech. 2004. Anaerobic ammonium oxidation in an estuarine
sediment. Aquat. Microb. Ecol. 36: 293-304.

124

Schubert C.J., Durisch-Kaiser E., Wehrli B., Thamdrup B., Lam P., and M.M.
Kuypers. 2006. Anaerobic ammonium oxidation in a tropical freshwater system (Lake
Tanganyika). Environ. Microbiol. 8:1857-1863.

Strous M., Kuenen J.G., and M.S.M. Jetten. 1999. Key physiology of anaerobic
ammonium oxidation. Appl. Environ. Microbiol. 65:3248-3250.

Thamdrup B., and T. Dalsgaard. 2002. Production of N2 through anaerobic ammonium
oxidation coupled to nitrate reduction in marine sediments. Appl Environ Microbiol
68:1312-1318.

Trimmer M., Nicholls J.C., and B. Deflandre. 2003. Anaerobic ammonium oxidation

measured in sediments along the Thames estuary, United Kingdom. Appl. Environ.
Microbiol. 69:6447-6454.

Tsushima I., Kindaichi T., and S. Okabe. 2007. Quantification of anaerobic
ammonium-oxidizing bacteria in enrichment cultures by real-time PCR. Water Res.
41:785-794.

Zhang Y., Xiao-Hong R., Op den Camp H.J.M., Smits T.J.M., Jetten M.S.M., and
M.C. Schmid. 2007. Diversity and abundance of aerobic and anaerobic ammonium-

oxidizing bacteria in freshwater sediments of the Xinyl River (China). Env. Microbiol.
9:2375-2382.

125

CHAPTER VI

CONCLUSIONS AND FUTURE PERSPECTIVES

At the time that this study was started, little was known about the distribution or
role of anaerobic ammonium oxidation in natural systems. The only environmental
evidence of anammox bacteria originated from anoxic water columns and sediments. The
screening primer set showed that anammox bacteria were present in shallow freshwater
systems, deep marine sediments, and Siberian permafrost. Their presence in a wide range
of environments suggested that anammox bacteria were not confined to strictly anaerobic,
stable habitats, which confirmed the initial hypothesis. The finding of the marine
anammox in freshwater systems showed that these Scalindua type anammox may be
found in concert with the classically “freshwater” Brocadia, Kuenenia, and Jettenia
anammox. Furthermore, identification of anammox bacteria in these habitats also
indicated bias against 16S rRNA gene amplification when universal primers are used.

The activity of anammox bacteria in the deep ocean was thought to dominate over
denitrification, with anammox responsible for greater than 80% of total N2 production.
The Cascadia Basin study showed that anammox was only responsible for about half of
the nitrogen loss from sediments at a depth of ~3000 m. Both denitrification and
anammox activity decreased with distance from shore, a proxy for sediment reactivity.
The lack of anammox dominance at low organic matter availability shows that
competition between denitrifiers and anammox at low sediment reactivity is still not well
understood. In fact, it appeared that ammonium availability was limiting anammox

activity in these deep sediments and not direct competition for nitrite. These findings did

126

not wholly disprove the initial hypothesis that anammox would dominate in these deep
sea sediments. Rather, the full testing of this hypothesis may require sampling of yet
deeper marine sediments, the abysmal plains, which constitute the majority of the world’s
oceans.

The development of the quantitative PCR method allowed for the first direct
correlation between anammox activity and abundance in marine sediments. The strong
correlation between q-PCR determined numbers and 15N isotope pairing measures of
activity indicates a strong reliance of activity on the size of the anammox community.
The vertical distribution of anammox down to 20 cm below sediment surface indicated
that survival is possible in locations that do not receive stable oxidized nitrogen inputs,
which corroborated with the alternative hypothesis. The presence of significant anammox
abundances in these zones further supported the conjecture from the survey study, that
the bacteria is present in environments not well suited, at least in terms of the current
knowledge, for maintenance of the anammox reaction. The vertical distribution showed
that these deep sea sediments were likely well-mixed due to macrofauna] bioturbation,
which created transient porewater channels for N03' and N02' movement into the anoxic
sediment.

The distribution of freshwater anammox in flooded agricultural sediments is
among the least understood areas of current anammox research. Large, unexplained
losses of added fertilizer N could not be accounted for in a previous study of taro fields
which suggested that anammox may be a significant N sink. The combination of q-PCR
primers allowed for the enumeration of both the “freshwater” and “marine” anammox

bacteria. The two magnitude lower abundance of anammox in the taro cropping system

127

suggests that they are a less dominate member of the microbial community, as compared
to the marine sediments from Cascadia Basin. However, the range of estimated anammox
activities based on previous measures indicate that anammox may be responsible for N
losses between 1.8 to 763.4 kg N ha'1 yr'l in taro fields, a potentially substantial N loss
where monthly fertilizer N applications are approximately 400 kg N ha'1 month'1 during a
nine-month cycle. The high numbers of anammox in the Hanalei River sediment indicate
that anammox may be a substantial N sink in this river which is likely loaded with runoff
from the surrounding taro patches. As a subject of future study, the overall importance of
the anammox reaction in these sediments will require confirmation using 15N isotope
pairing experiments.

Sediment microbial community structure did not exhibit any correlations with
geographic position, marine basin, depth below water surface, or water temperature
which disproved both tested hypotheses. The high proportion of highly divergent
sequences from the database demonstrated the depth obtained using 168 rRN A gene
pyrosequencing. The majority of these highly divergent sequences were within the rare
members at each site, suggesting that they are examples of microbial endemism and not
transients. The surprising lack of any cosmopolitan clusters, when grouped at 5%, would
appear to challenge the Baas-Becking hypothesis that “everything is everywhere, but the
environment selects”. The first part of this postulate contends that microbes are
distributed universally, in which we would expect to find some bacteria present in all of
our samples that share a commonality, e.g. they are all marine sediments. However, the
“but the environment selects” contends with niche differentiation and proliferation of

bacteria most suited to a particular environment. Clusters found at high abundance in

128

some sediments were not represented in other samples. This suggested that some
unidentified environmental factor was selecting for these prolific bacteria, despite that all
samples were from marine habitats. The high abundance of singletons suggested a very
high microbial diversity, though their functional importance to sediment nutrient cycling
is unknown.

The Florida Bay sediment samples provided an unique opportunity to investigate
the changes in microbial community structure between a high phosphorus and low
phosphorus site at long-terrn ecological research sites. The correlation of changes in
microbial relative abundances to past mesocosm and field studies that targeted certain
microbial groups showed that predation appeared to impact microbial community
structure among these two sites. While phosphorus availability likely determined the
coarse community structure of these sites, the relative abundances suggested that
phosphorus control likely led to biased predation on the most abundant members, which
allowed for a more robust rare biosphere illustrated by the highest diversity estimates at
the phosphorus impacted site. However, the true impact of predation on maintaining high

diversity is largely unknown and warrants further study.

Future Perspectives
As stated above, the importance of anammox in the abysmal depths of the world’s
oceans is currently unknown. Based on current evidence it is unlikely that anammox
dominates in these sediments due to substrate limitation. However, refinement of the
marine nitrogen budget requires that these questions be answered, particularly in the

context of oceanic productivity and its influence on carbon sequestration. While

129

anammox cell numbers are known to be linked to overall anammox activities, it is
unknown what the in-situ influences of N02", NO3', and NH4+ concentrations have on the
competition between denitrification and anammox activity on a per-cell basis. Once these
influences are known it may be possible to directly link porewater measures and q-PCR
determined anammox numbers to an estimated N2 production. As of now, only a broad
range of activities can be inferred from estimated cell counts.

Research into the importance of anammox in natural freshwater systems, flooded
agricultural fields, and terrestrial habitats is the future of anammox research. Nothing is
known regarding the activity of anammox in terrestrial systems and very little research
has been done in freshwater habitats. Early evidence has suggested that anammox plays a
small role in these systems, but the lack of coverage indicates that much more research is
necessary to cover a range of biogeochemical conditions found in the environment. In
particular, substantial anammox losses in agricultural systems would impact fertilization
schemes in order to reduce both nitrogen losses and monetary expenses.

As large scale sequencing efforts are undertaken in a variety of environments, our
knowledge of microbial diversity is expanding at an exponential rate. However,
uncovering the underlying causes of changes in community structure is a more
problematic issue. Future studies require the collection of large amounts of metadata,
such as porewater nutrient analysis, redox status, q-PCR determined cell counts, and
other indicators that can be correlated to relative abundances. Mesocosm studies can be
utilized to determine the influence of factors such as nutrient availability on the
composition and abundance of the “rare biosphere” and provide insight into the impact of

predation on total community diversity. One of the more difficult questions is whether the

130

rare members are functionally active and play a role in the sediment nutrient cycle. A
positive result would indicate that the current depth of sampling using pyrosequencing is
still not adequate in order to distinguish microbial community composition among sites.
On the other hand, if the rare members are not physiologically important then the current
sampling depth is not revealing any additional useful information. These are some of the
questions that must be addressed in the near future as sequencing capabilities continue to

advance in order to manage future research expenditures while maximizing information,

131

APPENDIX
CAN ONICAL CORRESPONDENCE ANALYSIS FOR
LINKING ENVIRONMENTAL VARIABLES WITH MICROBIAL

COMMUNITY STRUCTURE

As data rich methods such as microbial community pyrosequencing are
implemented in studies where biogeochemical parameters are also measured, it is
important for microbial ecologists to begin correlating these two datasets. Depending on
the goal of the study, these comparisons can be made at the level of the whole
community, phyla, or specific genera. The resounding question that needs to be addressed
concerns the statistical method used for comparisons. Classical methods utilized in
smaller clone library studies, such as principal correspondence analysis (PCA) are not
suited to the structure of pyrosequencing data sets. Abundant singletons, sequences found
only in one site, can account for as much as 70% of the total data set. As such, the
absence of operational taxonomic units (OTUs) creates a problem with PCA. Abundance-
based distributions are also not uniform, with a significant tail that reflects the number of
singletons in the dataset (Figure A-l). As such, the statistical method used for this type of
analysis must be able to overcome these issues.

Canonical correspondence analysis (CCA) was first developed as a tool to
investigate macrofauna] species-environment relationships. This was in order to
overcome a matrix of species data that is characterized by a prevalence of zeros and a

high ratio of species to samples (Clarke & Warwick, 1994). Additionally, in studies

132

where a specific environmental gradient is not sampled, the distribution of species would
be expected to take on a unimodal response. These hump-shaped unimodal distributions
cause problems in methods that assume linear response curves. However, canonical
correspondence analysis is not sensitive to this distribution (ter Braak, 1986, 1994). Other
canonical methods can be used where the species response to environmental variables
exhibits a linear response such as redundancy analysis (Jongman et al., 1995), a
constrained form of PCA, which is especially suited when sampling along environmental
gradients, even if these gradients are short.

Canonical correspondence analysis is also a robust method when species exhibit
bimodal ranges and unequal distributions. Additionally, it is well adapted for tables of
cluster-based abundances, deals well with species absence, and is sensitive to rare species
only when there is a small number of species. Specific OTUs can be identified that
respond to particular environmental variables (Ramette, 2007). In particular, CCA selects
ordination axes that are linear combinations of environmental variables. The best weights
for each environmental variable are then chosen as to maximize the dispersion of the
species scores. In CCA plots three items are displayed: species points, sample points, and
environmental arrows.

Canonical correspondence analysis was implemented in the R environment using
the VEGAN package. Tests of model confidence were based on the AN OVA permutation
test (anova.cca) and CCA permutation test (permutest.cca). The data structure for this
analysis consisted of over 290,000 168 rRNA gene sequences clustered at the 95%
identity level using the Ribosomal Database Project Pyrosequencing Pipeline complete

linkage clustering tool. The results show that the selected environmental variables

133

Available P, pH, SOC, and TN explained only 36% of the variability in the site
ordination. This is the underlying foundation of CCA in contrast to PCA. Principal
component analysis and other techniques perform ordination first on the community data
and then attempt to secondarily relate the ordination to the environmental factors. Instead,
CCA ordinates the community data based on multiple regressions on the relationships to
the environmental data. Simply, it uses the environment to constrain the ordination of the
community data. However, the choice of appropriate environmental variables has a
substantial effect on the ordination of the community data. In the case of this study, the
low percentage of variability explained by CCA is most assuredly due to the choice of
explanatory environmental variables, which were also correlated to each other. In order to
overcome this problem it is necessary to obtain sufficient sample metadata as to allow
insightful choices into what environmental variables should be included in the analysis.
Inclusion of variables that would not be expected to impact ordination introduce noise
into process and diminish the significance of the results (Mche, 1997).

Stepwise linear regression was used to further partition out the responses of
certain phyla to environmental variables. The interpretation of these results must be taken
with caution. Linear regression is meant to be used on continuous data, of which most
ecological data is not. Regardless, the absolute values of the regression coefficients and
statistical significance of their inclusion in a regression model can be used to infer their
overall influence in the distribution of particular phyla or other selected phylogenetic
level. However, as with any multivariate statistical method, artificial axes, clusters, and
variables do not necessarily correspond to actual biological or ecological conditions

(Ramette, 2007).

134

The full R script used for this analysis follows:

#The Ade4 library is used tmnerate plots using the following script:
library(adegenet)

library(ade4)

# Switch screens only following the enter key

par(ask=T)

# 10 mercenttxt is the cluster file previously arcsin transformed in Excel
fau<-read.table(' l 0percent.txt',header=TRUE,sep=',',na.strings="NA")
# nutrient.txt is the environmental variable file
mil<-read.table(‘nutrient.txt',header=TRUE,sep=',',na.strings="NA")
# Log transform the nutrient data

lmil<-log(mil+1)

# Ensure that the data is still in the proper format

summary(lmil)

summary(fau)

list(lmil)

list(fau)

# Perform ccaLthe cluster file is listed first

ivl <- cca(fau,lmil, scan = FALSE)

plot(iv1)

# Projections of inertia axes on PCAIV axes

s.corcircle(iv1 $as)

135

# Species positions

 

s.label(iv1$cl, 2, 1, clab = 0.5, xlim = c(-4,4))

# Sites positions at the weighted mean of present smcies
s.label(iv1$ls, 2, l, clab = 0.5, cpoi = l, add.p = TRUE)

# Prediction of the positions by regression on environmental variables
s.match(iv1$ls, ivl$li, 2, 1, clab = 0.5)

# Canonical weights givingunit variance combirgrtions
s.arrow(iv1 $fa)

# Position of species by averaging

s.label(ivl$ll, 2, 1, clab = 0.5, cpoi = 1.5,add.plot=TRUE)
s.label(iv1$co, 2, l, clab=0.5,add.plot = TRUE)
s.distri(ivl$ll, fan, 2, 1, cell = 0, csta = 0.33)
s.label(iv1$co, 2, l, clab = 0.75, add.plot = TRUE)

# Coherence between weights and correlations

par(mfrow = c(1,2))

s.corcircle(iv1$cor, 2, 1)

s.arrow(iv1$fa, 2, 1)

par(mfrow = c(1,1))

#The vegan library is used twmfle the statistics:
library(vegan)

par(ask=T)

# 3percent.csv is the cluster file which has been previously arcsin transformed

# nutrient.txt is the environmental variables file

 

136

fau<-read.table('3percent.csv',header=TRUE,sep=',',row.names=l )
mil<-read.table(‘nutrient.txt',header=TRUE,sep=',',na.strings="NA")
# Check the integrity of the input file:

str(fau)

# Transform the environmental variables mther transformations may be more correct,
depending on the data structure

lmil<-log(mil+l)

# Checlghat the data structure is still correct

summary(fau)

list(lmil)

list(fau)

# Perform canonical correspondence an_alysis on the two data sets

ivl <- cca(fau,mil, scan = FALSE)

ccal .plot<-plot(ivl )

text(cca1 .plot,what='sites',col="black",pch=2,cex=0.7,font=2)

goodness(iv1)

goodness(ivl, summ = TRUE)iv2 <- cca(fau$cluster
~mil$soc+mil$Ph+mil$biomass)vare.ccaplot(vare.cca)

# Anova and CCA permutation tests are used to test the validity of the model
anova(iv1, alpha=0.05, beta=0.01, step=100, perm.max=10000)
permutest.cca(iv1, permutations=10000, model=c("reduced"))

# Returns the proportion of inertia accounted for by the axis

137

99 93

goodness(iv1,display=c(“sites , species”), model = c("CA"), statistic =
c("explained"),summarize = FALSE)

ivl $CCA$alias

spec <- iv1$CCA$v

# Run regression of SOC vs cluster linear model regression
sm <- lm(percent~soc)

# How good is the model?

summary(sm)

# Perform correlations and stepwise regressions of interesting phyla versus environmental
variables

cor(percent, soc)

cor(percent,soc)"2

#Wh_at does the relationship log]; like?

plot(percent,soc)

# Plot the linear model

abline(sm$coeff)

title(main="Linear Model Plot")

# Returns the slgae and intercept of the regession line
list(smScoeft)

# Plot the model residuals

plot(sm$resid,)

abline (0,0)

title(main="Model residuals")

138

#LOCAL REGRESSIONS

#local.east <- loess(percent ~ soc)
#plot(soc,percent)
#points(soc,local.east$fitted,col="orange",pch=4)

#summary(local.east)

139

REFERENCES

Clarke K.R., and RM. Warwick. 1994. Change in Marine Communities: an Approach
to Statistical Analysis and Interpretation, Plymouth Marine Laboratory, Natural
Environment Research Council, UK.

Jongman R.H.G., Ter Braak C.J.F., and O.F.R. Van Tongeren. 1995. Data Analysis
in Community and Landscape Ecology, Cambridge University Press, UK.

McCune B. 1997. Influence of noisy environmental data on canonical correspondence
analysis. Ecology 78:2617-2623.

Ramette A. 2007. Multivariate analyses in microbial ecology. FEMS Microbiol. Ecol.
62:142-160.

ter Braak C.J.F. 1986. Canonical correspondence analysis: a new eigenvector technique
for multivariate direct gradient analysis. Ecology 67:1167-1179.

ter Braak C.J.F. 1994. Canonical community ordination. Part I: Basic theory and linear
methods. Ecoscience 1: 127-140.

140

 

”'illlllllllllillillilllillilllllll“