MICROBIAL SUCCESSION ON THE LAKE STURGEON EGG SURFACE:
MECHANISMS SHAPING THE MICROBIAL COMMUNITY ASSEMBLY DURING
SUCCESSION AND THE EFFECT OF MICROBIAL SUCCESSIONAL PROCESSES ON
HOST LIFE HISTORY TRAITS
By
Masanori Fujimoto

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Microbiology and Molecular Genetics
Ecology, Evolutionary Biology and Behavior
2012

ABSTRACT
MICROBIAL SUCCESSION ON THE LAKE STURGEON EGG SURFACE:
MECHANISMS SHAPING THE MICROBIAL COMMUNITY ASSEMBLY DURING
SUCCESSION AND THE EFFECT OF MICROBIAL SUCCESSIONAL PROCESSES
ON HOST LIFE HISTORY TRAITS
By
Masanori Fujimoto
Microbial community assemblages have been studied in a variety of hosts and
environments. However, to date, most of the studies conducted on microbial community
structure have been observational in nature. As a result, the underlying mechanisms shaping
microbial community assembly at a given place and time remain largely unknown. In this study,
we were particularly interested in understanding how a microbial community develops on animal
hosts. As soon as eggs or neonatal organisms are exposed to environments, microbes colonize
surfaces of eggs or the epithelium of host tissues and establish a microbial community. Such a
process involving initial colonization and subsequent sequential changes in species composition
is called “succession”. Microbial succession is a complex process with the number of different
factors involved including initial stochastic arrival of microbes at an open space via dispersal,
attachment / colonization at the space, subsequent community sorting via adaptation, and
continuous dispersal from neighboring spaces. In this study, we examined microbial succession
on the egg surface of the Lake Sturgeon, a threatened fish species inhabiting in the Great Lakes.
We sought to understand the role of both host factors (e.g. innate immunity, egg chemistry) and
various environmental factors (e.g. aquatic microbes, stream flow rate and temperature) in
influencing the formation of microbial communities on the egg surfaces over the course of the
egg incubation period. We also sought to evaluate the effect of different microbial successional
processes on host life history traits, including egg mortality and larval size at hatch. These topics

were important for this system because dams constructed in the Lake Sturgeon’s spawning
streams can alter environmental factors such as aquatic microbes, stream flow rate, and
temperature, which may in turn affect the life history of the sturgeon.
To achieve these objectives, we adopted an integrative approach, which relied on
manipulation of environmental factors including aquatic microbial community, aquatic microbial
quantity, stream flow rate, and temperature. We also employed a combination of various
analytical techniques, including 16S rRNA gene pyrosequencing, 16S rRNA based T-RFLP, 16S
rRNA clone library, 16S rRNA based quantitative PCR, light and fluorescence microscopy, and
culture methods. We found that egg microbial communities were distinct from source water
microbial communities. Host eggs shaped egg-associated microbial communities within 60
minutes of fertilization, selecting for and against certain microbial species. In addition, the egg
surface microbial communities were not constant but rather dynamic, as we observed directional
changes of microbial communities along with egg developmental stages. Egg-associated
microbial communities also varied with the environmental variables they were exposed to during
incubation, including temperature, flow rate, and aquatic microbial community. These
differences in the egg-associated microbial community composition affected host life history
traits including egg mortality and larvae size at hatch. We also identified a key set of microbial
species that significantly improved egg survival and could be used for probiotic treatment in this
threatened fish species in the future. This study was the first microbial succession study
conducted on fish eggs. Our results highlight the complexity of host-microbe-environment
interactions. This study has implications for managing threatened host populations such as the
Lake Sturgeon inhabiting human-altered rivers, since it demonstrates the potential effect of dams
(which alter aquatic microbes and temperature) on downstream host-microbe interactions.

Copyright by
MASANORI FUJIMOTO
2012

This dissertation is dedicated to my parents and to my friend Wilhelmina Yonkman

v

ACKNOWLEDGEMENTS

I thank Dr. Terry Marsh for having me in his lab for the last 8 years ever since I was still
in the Engineering program. I thank him for introducing me to this exciting Lake Sturgeon
project and many other projects, all of which helped me to learn and acquire knowledge, skills
and life lessons. He taught me all of the microbiology and molecular techniques which were used
to conduct this project side by side lessons which will remain with me for the rest of my life. I
thank him for always being available for me. He always paid attention to detail and provided
many different perspectives to this project. He adopted me from the Engineering and nurtured
me so someday I can walk alone. I really appreciate the work he has done for me, and without
question I could not do this project without him.
I thank Dr. Kim Scribner for giving me this incredible opportunity and journey of being a
part of the Lake Sturgeon project. He treated me as one of his own and he molded me as an
ecologist and evolutionary biologist. He guided me to a direction of new growth in my
knowledge base and career. I really appreciate his passion, curiosity and honesty in science. He
taught me the importance of hard work that is necessary to conduct science. Staying in a field
station in Onaway, Michigan with him and his crew was one of the best moments in my life
which I will always remember.
I thank Dr. Ned Walker for his input into this project during committee meetings. He
contributed greatly to my growing understanding of microbial ecology principles. I really thank
him for his support throughout my Ph.D. program not only during the committee meeting but
also whenever he saw me in the hallway. He encouraged me any time he had the chance and

vi

cared about me, always asking me if I was doing okay. Interactions with him made me feel
positive about my work and about myself.
I thank Dr. Todd Ciche for his support at committee meetings and for bringing a different
perspective and knowledge to the table, specifically about host-microbe interactions. He also
introduced me to new information which is relevant to this project. He was always available for
me and every time I visited his office, he welcomed me and took time for me. I really thank him
for that.
I also thank Dr. Bob Hausinger who gave me a great deal of advice about my academic
program and also gave me an opportunity to teach the MMG 302 class. I also thank him for
helping me to get funding including a travel grant to attend a conference and a Dissertation
Completion Fellowship, both of which were much appreciated and assisted me in completing my
program.
I also thank my field associates: John Bauman, Katy Jay, Ryan Hastings, and Jared
Homola who spent two years with me in the field, and Jason Lorenz, Matt, Aaron Orr, Nikki,
Ryan Young, and Bill, who spent one year with me in the field. We could not keep this project
going without their help and their tireless work. I also thank my fellow graduate students in the
field- Jamie Crossman, Ben Rook and Kari Dammerman. I am grateful for the DNR personnel
who assisted with this work, especially Dr. Ed Baker, who has been a great help in this project
not only in the field but also bringing his diverse experiences to all aspects of the project. I also
thank Jim Holser from the Department of Natural Resources (DNR) who assisted us greatly with
our work in the hatchery. I also thank our local friends at the field Mel for provided housing for
us during the field season and Gary for diving in deep holes of the stream with his wet suit. I
also appreciate the members of the Sturgeon for Tomorrow organization for their tireless work in

vii

protecting the sturgeon throughout the spawning season, and for their financial support for our
project.
I am indebted to my former and current ROME lab members: Sang Hoon Kim, Christina
Harzman, Jen Mayrberger, Fan Yang, Natasha Isaacs, Joe Duris, Lisa Fogarty and Britton
Hildebrandt for their help and support during my graduate studies. I specifically thank Dr. Sang
Hoon Kim for his support in teaching me lab work skills and providing me resources that I
needed to become a good microbiologist. I also thank my undergraduate assistants in MMGBrian Lovett and Paul Nirenberg for assisting with antagonistic interaction work and Emily
Cannell in Fisheries and Wildlife for helping me with ImageJ analysis. I am grateful to Dr.
Mohamed Faisal and Dr. Tom Loch for providing resources for this project, specifically fish
pathogens (which were one of the key components in Chapter 6). I also thank Dr. Tom Voice
who was my previous mentor in Environmental Engineering for his support in guiding me in the
right direction in the earlier stages of my PhD program. I also thank Dr. Tom Whittam in MMG
for his advice in the earlier stages of my PhD program.
I am indebted to the funding sources of this project including the Department of Natural
Resources (DNR), the United States Geological Survey (USGS), Sturgeon for Tomorrow, Great
Lakes Fisheries Trust, Sustainable Michigan Endowed Project, and Michigan State University
(MSU). I especially thank MSU for not only providing internal funding for this project, but also
supporting my graduate studies, including via a travel grant to attend the Ecological Society of
America (ESA) conference and a Dissertation Completion Fellowship, which allowed me to
fully focus on this project during the critical final stages.
I also thank Lyman Briggs College for giving me opportunity to teach general chemistry
for 5 straight years. Specifically, I thank Dr. Maxine Davis, who gave me an opportunity to teach

viii

her classes and always trusted me and encouraged my growth as a teacher. I also thank my
colleagues in Lyman Briggs who helped me grow through collaboration. I also thank all of my
students. I learned so much from teaching and will carry many rewarding experiences I had in
the classroom with me in my future endeavors.
I thank my family friend Wilhelmina Yonkman for helping me throughout my life in
graduate school. I appreciated her friendship and her positive energy that was contagious. I also
thank her son and daughter-in-law Leonard and Alice Yonkman for their friendship. They treated
my wife and I as their own kids and we have always appreciated that.
I thank my friends in East Lansing, at MSU, in the Microbiology and Molecular Genetics
department specifically Maris Laivenieks, in the Fisheries and Wildlife department and at the
Center for Systems Integration and Sustainability for their friendship. I also thank my friends
over in Japan, specifically Naoki Okumura, Osamu Sakamoto, and Tomo Nagai for their support.
Lastly, I thank my family for their support. I thank my father-in-law, mother-in-law and
brothers-in-law for their love and support. I thank my brother and sister in Japan for taking care
of my parents when I have been away from home. I thank my mom and my dad for allowing me
to come to the United States and giving me an opportunity to explore a new life here. Finally, I
thank my wife Vanessa for her unconditional love over many years. I also thank my as yet to
arrive baby for encouraging me throughout this last stretch of my dissertation. His or her
presence in the womb has kept me going over the course of these last few months and I look
forward to the day we meet.

ix

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................................................... xii
LIST OF FIGURES ................................................................................................................... xiii
KEY TO SYMBOLS OR ABBREVIATIONS ........................................................................ xvi
CHAPTER 1: INTRODUCTION................................................................................................ 1
Global scope of this study ................................................................................................... 1
Overview of chapters .......................................................................................................... 5
References ........................................................................................................................... 8
CHAPTER 2: MICROBIAL COMMUNITY ASSEMBLY AND SUCCESSION ON LAKE
STURGEON EGG SURFACES AS A FUNCTION OF SIMULATED SPAWNING
STREAM FLOW RATE ............................................................................................................ 13
Abstract ............................................................................................................................. 13
Introduction ....................................................................................................................... 15
Methods............................................................................................................................. 17
Results ............................................................................................................................... 23
Discussion ......................................................................................................................... 37
Appendix ........................................................................................................................... 43
References ......................................................................................................................... 48
CHAPTER 3: THE EFFECT OF TEMPERATURE AND WATER TYPE ON THE EGG
SURFACE MICROBIAL COMMUNITY AND HOST LIFE HISTORY TRAITS OF THE
LAKE STURGEON .................................................................................................................... 54
Abstract ............................................................................................................................. 54
Introduction ....................................................................................................................... 56
Methods............................................................................................................................. 57
Results ............................................................................................................................... 63
Discussion ......................................................................................................................... 74
References ......................................................................................................................... 79
CHAPTER 4: THE EFFECT OF INOCULATION OF PUTATIVE SYMBIONTS ON
LAKE STURGEON EGG SURFACE MICROBIAL COMMUNITY ASSEMBLY AND
EGG MORTALITY.................................................................................................................... 83
Abstract ............................................................................................................................. 83
Introduction ....................................................................................................................... 85
x

Methods............................................................................................................................. 87
Results ............................................................................................................................... 93
Discussion ....................................................................................................................... 106
References ....................................................................................................................... 111
CHAPTER 5: THE RELATIVE IMPORTANCE OF REGIONAL DISPERSAL AND
LOCAL DETERMINISTIC PROCESSES IN SHAPING THE MICROBIAL
COMMUNITY ASSEMBLY ON THE EGG SURFACES OF THE LAKE STURGEON
(ACIPENSER FULVESCENS) ................................................................................................ 115
Abstract ........................................................................................................................... 115
Introduction ..................................................................................................................... 117
Methods........................................................................................................................... 120
Results ............................................................................................................................. 126
Discussion ....................................................................................................................... 147
References ....................................................................................................................... 152
CHAPTER 6: CHARACTERIZATION OF BACTERIAL ISOLATES FROM THE EGG
SURFACES OF LAKE STURGEON (ACIPENSER FULVESCENS) FOR
ANTAGONISTIC INTERACTIONS AND BIOFILM FORMING CAPABILITIES ...... 158
Abstract ........................................................................................................................... 158
Introduction ..................................................................................................................... 160
Methods........................................................................................................................... 162
Results ............................................................................................................................. 169
Discussion ....................................................................................................................... 183
References ....................................................................................................................... 188
CHAPTER 7: CONCLUDING REMARKS .......................................................................... 192

xi

LIST OF TABLES

Table 2.1. Bray-Curtis dissimilarity index matrix using TRFLP data summarized by day.......... 25
Table 2.2. Bray-Curtis dissimilarity matrix using the RDP classifier output at genus level ........ 32
Table 2.3. Estimated species evenness (E) using OTUs at 97% similarity cutoff ........................ 36
Table 3.1. 16S rRNA gene-based clone library for the two different treatments at two
different time points for the CE family. ........................................................................................ 69
Table 4.1. Comparison of stream water microbial community and stream water fertilized egg
microbial community at the phylum level using 454 pyrosequencing data................................ 101
Table 4.2. Comparison of water and egg microbial communities at the genus level using 454
pyrosequencing data.................................................................................................................... 103
Table 5.1. Comparison of microbial communities in different water types using 454
pyrosequencing analyzed at the phylum/class level. .................................................................. 129
Table 5.2. Comparison of microbial communities in different water types using 454
pyrosequencing analyzed at the genus level. .............................................................................. 130
Table 5.3. Comparison between the egg associated microbial communities and source water
microbial communities at 6 hours after fertilization using 454 pyrosequencing analyzed at
the phylum level. ......................................................................................................................... 136
Table 5.4. Comparison of the egg associated microbial communities and source water
microbial communities at 6 hours after fertilization using 454 pyrosequencing analyzed at
the genus level............................................................................................................................. 137
Table 5.5. Water treatment effect on egg microbial community plus temporal trend analyzed
using 454 pyrosequencing at genus level (RC family data). ...................................................... 139
Table 6.1. List of isolates used for antagonistic interactions. ..................................................... 165
Table 6.2. Antagonistic interactions among 25 egg isolates at two different temperatures ....... 171

xii

LIST OF FIGURES

Figure 2.1. Temporal and flow rate effects on microbial community composition during
microbial succession. .................................................................................................................... 24
Figure 2.2. Principal component score plots revealing the temporal clustering of microbial
community assemblages on egg surfaces. .................................................................................... 27
Figure 2.3. Principal component loading plots displaying temporal distributions of 25 major
microbial phylotypes (PT) associated with the egg surface.......................................................... 28
Figure 2.4. Microbial phylotype richness and time relationships during succession across
flow regimes (High ○, Low +, and Variable ∆). ......................................................................... 29
Figure 2.5. Characterization of the egg surface microbial community assembly using the
RDP classifier output at the phylum level (and class level for Proteobacteria)............................ 31
Figure 2.6. Characterization of the egg surface microbial community assembly using the
RDP classifier output at the genus level. ...................................................................................... 32
Figure 2.7. Principal component score plots using OTUs defined at 97% similarity cutoff. ....... 35
Figure 2.8. Microbial phylotype richness (97% OTUs rarefied at 6000 reads) and time
relationships during succession across flow regimes.................................................................... 36
Appendix Figure A.2.1. Examples of the association of certain microbial phylotypes with
certain egg developmental stages.................................................................................................. 44
Appendix Figure A.2.2. Rarefaction analysis for pyrosequencing data. ...................................... 45
Appendix Figure A.2.3. Jaccard index tree constructed using OTUs at 97% similarity cutoff. ... 46
Appendix Figure A.2.4. Ambient water temperature throughout the egg incubation periods...... 47
Figure 3.1. Principal component analysis (PCA) plot of TRFLP data of 72 samples from
different treatments. ...................................................................................................................... 64
Figure 3.2. Principal component analysis (PCA) plot depicting the effect of water treatment
on egg microbial community at warm temperature. ..................................................................... 66
Figure 3.3. Positive linear relationship between time (days post-fertilization) and microbial
quantity present on egg surfaces. .................................................................................................. 68
Figure 3.4. The average microbial quantity associated with eggs as measured using qPCR. ...... 70
Figure 3.5. The effect of water treatment on egg mortality. ......................................................... 71
xiii

Figure 3.6. The effect of temperature and water treatment on resource allocation. ..................... 73
Figure 4.1. A box plot showing the effect of the treatments on egg mortality. ............................ 94
Figure 4.2. Community assembly of unfertilized eggs (Unf) and eggs immediately after
being fertilized (AF). .................................................................................................................... 96
Figure 4.3. Convergence of the egg microbial community structure at Day 2 postfertilization. ................................................................................................................................... 97
Figure 4.4. Principal component analysis (PCA) analysis using TRFLP data from both egg
and water samples. ........................................................................................................................ 98
Figure 4.5. Microbial quantity on the egg surfaces of various treatments and controls at
different time points estimated using qPCR. .............................................................................. 100
Figure 4.6. Potential vertical transmission of Acidovorax sp. F19. ............................................ 104
Figure 4.7. Box plot depicting the effect of Acidovorax sp. F19 treatment on yolk sac area to
body area ratio of eggs. ............................................................................................................... 105
Figure 5.1. Schematic diagram of experimental design to manipulate initial inocula and
subsequent rate of dispersal. ....................................................................................................... 122
Figure 5.2. Direct microscopic counts for microbial density in each water type. ...................... 127
Figure 5.3. The effect of water type (aquatic microbial community) on the egg surface
microbial communities as measured using T-RFLP. .................................................................. 131
Figure 5.4. PCA plot showing that dispersal was dependent on water microbial density. ......... 132
Figure 5.5. PCA analysis on the relative importance of dispersal and deterministic processes
in community assembly as measured with T-RFLP. .................................................................. 134
Figure 5.6. Quantification of the egg associated microbes across different treatments using
qPCR. .......................................................................................................................................... 142
Figure 5.7. The effect of transfer from stream water to 0.2 µm filtered water on the egg
surface microbial quantity........................................................................................................... 143
Figure 5.8. The effect of rearing water type on yolk sac resource uses. ..................................... 145
Figure 5.9. A box plot showing the effect of fertilization and rearing environment on egg
mortality. ..................................................................................................................................... 146
Figure 6.1. Antagonistic interaction screening and confirmation using soft agar overlay
assay. ........................................................................................................................................... 167

xiv

Figure 6.2. Phylogenetic relationships among 92 sturgeon egg isolates inferred using the
Neighbor-Joining method. .......................................................................................................... 170
Figure 6.3. Summary of (a) lawn susceptibilities and (b) stamp aggressiveness among 25
sturgeon egg surface isolates at two temperature regimes .......................................................... 174
Figure 6.4. Reciprocal antagonistic interactions among 25 egg surface isolates when
positions are switched from stamp to lawn. ................................................................................ 176
Figure 6.5. Reduction of colony size due to both active growth inhibition by lawn isolate and
resource competition under two temperature regimes. ............................................................... 177
Figure 6.6. Antagonistic interactions against 6 known fish pathogens by the 4 most
aggressive isolates identified on the sturgeon egg surface. ........................................................ 179
Figure 6.7. Quantification of biofilm formation with crystal violet in a microtiter plate assay
under different media .................................................................................................................. 181
Figure 6.8. Quantification of planktonic growth in different nutrient broth with OD600 in a
microtitier plate assay. ................................................................................................................ 182

xv

KEY TO SYMBOLS OR ABBREVIATIONS

16 S rRNA………………….a component of the 30S small subunit of prokaryotic ribosomes
Ct…………………………………………….…………………………………cycle threshold
OTU………………………………………………………………..operational taxonomic unit
PBS………………………………………………………………….phosphate buffered saline
PCA……………………………………………………………..principal component analysis
PCR…………………………………...…………………………….polymerase chain reaction
PT…………………………………………………………………………...……….phylotype
qPCR…………………………………………………..quantitative polymerase chain reaction
RDP………………………………………………………………Ribosomal Database Project
RFU…………………………………………….……………………relative fluorescence unit
rRNA…………………………………………….…………………ribosomal ribonucleic acid
TRFLP……………………………………terminal restriction fragment length polymorphism

xvi

CHAPTER 1: INTRODUCTION
Global scope of this study
In this present study, we sought to answer fundamental ecological questions, specifically
understanding ecological processes shaping microbial community assembly at any given time
and place. Microbial community assemblages have been studied in a variety of environments
ranging from marine [1-2], lake [3-4], soils [5-6], plants [7-8], animal gut [9], and human gut
[10]. The variability that can be observed in the structure of microbial communities across
diverse hosts and environments is particularly fascinating and has significant implications for
more broadly understanding fundamental ecological roles of microbes. However, to date, most
of the studies conducted on microbial community structure have been observational in nature. As
a result, the underlying mechanisms shaping microbial community assembly at a given place and
time remain largely unknown, as does the role of microbes in the greater ecological system.
One topic of interest in the study of microbial ecology involves understanding how a
microbial community develops on animal hosts. Because microbes are ubiquitous [11], their
association with a variety of different hosts is inevitable. Microbes utilize host resources and
hosts in turn can be benefited or harmed by the presence of microbes [12]. This interaction plays
a significant role in both a host’s life history and a microbe’s life cycle, yet much remains
unknown regarding the exact nature of ecological processes governing microbe-host interactions.
Microbe-host interactions start from the beginning of a host’s life. Embryos of animals
are particularly vulnerable to microbial challenges since their immunity is not fully developed
[13-14]. Therefore, both innate immunity [13, 15] and adaptive immunity [14, 16] can be
provisioned to embryos by hosts through various different mechanisms. Animals have evolved to
protect embryos from microbes via formation of an egg case [13] and a placenta [14]. As soon as

1

animals are born and exposed to their outside environments, the eggs or neonatal organisms
come into contact with microbes [17-18]. Microbes colonize the surfaces of eggs or epithelium
of host tissues and establish a microbial community.
The processes of initial colonization of hosts by collections of microbes and subsequent
sequential changes in microbial community structure on hosts have been documented [18] and
such a process is called “succession”. Succession has been more broadly investigated across the
field of ecology for many years, mainly using plant systems [19]. However, microbial succession
has been recently studied in various host animals [18, 20-21], host plants [7], and natural [22-25]
and artificial [26-28] environments. Results from these studies suggest that microbial succession
is a complex process [7, 18, 22, 29]. A number of different factors are involved in microbial
succession including initial stochastic arrival at an open space via dispersal from neighboring
spaces, attachment-colonization at the space, subsequent community sorting via adaptation, and
continuous dispersal from neighboring spaces [29]. Microbes serve as ideal models for studying
succession because the microbial succession process can be effectively controlled in an
experimental setting, thus allowing for novel information to be learned about factors affecting
the succession process.
In this study, we examined microbial succession on the egg surface of Lake Sturgeon
(Acipenser fulvescens), a threatened fish species inhabiting in the Great Lakes. The Lake
Sturgeon is one of the 26 sturgeon species inhabiting freshwater [30]. The species is unique in
that it has maintained its ancestral morphological form since the Lower Jurassic period [30].
Lake Sturgeon populations have decreased drastically over the past 100 years due to
anthropogenic activities such as overfishing and dam construction [31]. One such population is
the Black Lake population in Michigan [31-32]. Spawning habitats have been altered since the

2

construction of Kleber dam in 1949 on the Upper Black River, which is the sole spawning stream
for the Black Lake population. Despite recent restoration efforts, natural recruitment is limited,
which is likely attributed to high egg mortality [33].
Microbial succession on eggs of the Lake Sturgeon is likely a complex process that is
worthy of focused study. Eggs are fertilized in a stream as soon as male and female adults release
gametes. A female is usually surrounded by multiple males and releases over 500,000 eggs per
spawning season [30]. The water activated eggs develop an adhesive quality [34] so that they can
adhere to benthic substrates such as gravel and sand [30]. During this fertilization process,
microbes that are drifting in a stream collide with eggs and adhere to the sticky egg surfaces.
This process is a stochastic process, since the water microbial community varies temporally and
spatially [24, 35-36], and microbes drifting in stream water have no control over their
movements. After the initial stochastic collision, the microbial community on egg surfaces are
likely selected by local deterministic processes including adhesion [37-38], antimicrobial
activities of eggs [15, 39], chemicals that eggs excrete during embryogenesis [40-41], and
interspecific competition among microbes [42-46], while microbes in stream water continuously
colonize egg surfaces via passive dispersal mediated by water flow.
A number of other factors may potentially affect such successional process such as
stream flow rates [25], temperature [24], and structure of the aquatic microbial community [47].
These factors are important to consider in this system because dams in the Lake Sturgeon’s
spawning streams can alter environmental factors such as aquatic microbes [36], stream flow rate
[48] and temperature [49], but the effects of such changes on microbe communities and egg
survival are unknown. Microbe-host interactions are long-term in nature and have in many cases
coevolved to the point of reaching a fine balance. The perturbation of these interactions as a

3

result of anthropogenic activities or artificial substances may cause an observable disturbance to
the system that is worthy of further study. This is of particular concern for the Lake Sturgeon
host when considering the observed high egg mortality rate and the potential for microbe-host
interactions to play a role in influencing such mortality, a topic that has not been studied to date.
In this dissertation work, we studied microbial succession on the Lake Sturgeon egg
surfaces over the course of their incubation period and under different environmental conditions.
The overarching goal of this study was to characterize the egg surface microbial community
assembly during succession and to examine how both host factors (e.g. innate immunity, egg
chemistry) and environmental factors (e.g. aquatic microbes, stream flow rate and temperature) affect
the microbial succession process and subsequently influence host life history traits (e.g. egg
mortality and larval size at hatch). It was also our goal to identify putative symbionts for the egg
of the Lake Sturgeon to be potentially used in the future for probiotic treatment of eggs in
hatcheries. We believed that characterizing microbial community assembly on the egg surface
during incubation would help illuminate the potential but as yet unexplored causal relationship
between microbes and egg mortality.
To achieve these goals, we took an integrative approach, which relied on experimental
manipulation of environmental factors including aquatic microbial community, aquatic microbial
quantity, stream flow rate, and temperature while rearing Lake Sturgeon eggs in a streamside
hatchery. We then monitored the changes in microbial communities across these treatments and
over the course of the egg developmental stages using a combination of various analytical
techniques, including 16S rRNA gene pyrosequencing, 16S rRNA based T-RFLP, 16S rRNA
clone library, 16S rRNA based quantitative PCR, light and fluorescence microscopy, and culture
methods.

4

Overview of chapters
Chapter 2
The primary objective of this chapter was to document microbial succession (initial
colonization and subsequent sequential changes in microbial community assembly) on the
surface of the Lake Sturgeon eggs and to characterize the egg associated microbial community
assembly using next generation sequencing. We also examined the effect of simulated stream
flow rate on the microbial community assembly on the egg surfaces. The effects of natural and
dammed stream flow rates on microbial succession were also discussed.

Chapter 3
In this chapter, we analyzed the effect of environmental variables, including temperature
and aquatic microbial composition, on microbial succession on the Lake Sturgeon egg surfaces.
We also examined the effect of such environmental variables on subsequent life history traits of
the Lake Sturgeon egg host, including egg mortality and larvae size at hatch. We analyzed these
effects using an experimental design with controlled treatments, including two temperature
ranges (warm and cold) and two water types (stream water and UV treated stream water). The
changes in egg microbial community during succession and across treatments were analyzed,
and the correlation between the microbial community assembly and the egg mortality was
discussed.

Chapter 4
In this chapter, Lake Sturgeon eggs were fertilized with a putative symbiont and putative
pathogen that were previously isolated from egg surfaces. After initial inoculation, successional

5

microbial community changes on the egg surfaces were examined. The effect of inoculation of
the putative symbiont on egg mortality and other life history traits (including larvae size at hatch)
was also studied. The potential application of the putative symbiont isolate for probiotic
treatment in the Lake Sturgeon hatchery was also discussed.

Chapter 5
In this chapter, the relative importance of dispersal (the effect of aquatic microbes) and
local deterministic effects (host egg effects) in shaping the egg associated microbial community
was studied by manipulating the aquatic microbial community composition and quantity
throughout embryogenesis. Eggs were fertilized and reared in three different water types (stream
water, UV treated stream water, and 0.2 µm filtered stream water) which had different aquatic
microbial community compositions and concentrations. We hypothesized that if the dispersal is
dominant processes in shaping the egg associated microbial community, egg microbial
community would converge with the source water microbial communities. We characterized
both the egg surface microbial communities and the aquatic microbial communities across
treatments and throughout embryogenesis to elucidate the relative roles of dispersal of aquatic
microbes and local egg-related host effects on microbial community assembly.

Chapter 6
In this chapter, 25 microbial strains isolated from Lake Sturgeon egg surfaces were
characterized for both antagonistic microbial interactions and biofilm forming capabilities. The
soft agar overlay technique and the crystal violet staining technique were used to evaluate
antagonistic microbial interactions and biofilm forming capability, respectively. This experiment

6

was performed to identify a microbe that interacts with fish pathogens antagonistically and thus
one that would be a good candidate for probiotic treatment of Lake Sturgeon eggs. We also
assessed whether microbial interactions among isolates and biofilm formation of each isolate
help explain the changes in the egg microbial community structure observed in previous chapters.

Chapter 7
Findings from previous chapters were summarized. Mechanisms that govern the
formation and development of the egg surface microbial communities were discussed based on
the findings of the dissertation. Implications of our findings and future suggestions were also
provided.

7

References

8

References

1.

Huber JA, Mark Welch DB, Morrison HG, Huse SM, Neal PR, Butterfield DA, Sogin
ML (2007) Microbial Population Structures in the Deep Marine Biosphere. Science 318:
97-100

2.

Pommier T, Canback B, Riemann L, Bostrom KH, Simu K, Lundberg P, Tunlid A,
Hagstrom Å (2007) Global patterns of diversity and community structure in marine
bacterioplankton. Molecular Ecology 16: 867-880

3.

Van der Gucht K, Vandekerckhove T, Vloemans N, Cousin S, Muylaert K, Sabbe K,
Gillis M, Declerk S, De Meester L, Vyverman W (2005) Characterization of bacterial
communities in four freshwater lakes differing in nutrient load and food web structure.
FEMS Microbiology Ecology 53: 205-220

4.

Yannarell AC, Triplett EW (2005) Geographic and environmental sources of variation in
lake bacterial community composition. Appl Environ Microbiol 71: 227-239

5.

Fierer N, Jackson RB (2006) The diversity and biogeography of soil bacterial
communities. Proceedings of the National Academy of Sciences of the United States of
America 103: 626-631

6.

Garbeva P, Postma J, Van Veen JA, Van Elsas JD (2006) Effect of above-ground plant
species on soil microbial community structure and its impact on suppression of
Rhizoctonia solani AG3. Environmental Microbiology 8: 233-246

7.

Redford A, Fierer N (2009) Bacterial Succession on the Leaf Surface: A Novel System
for Studying Successional Dynamics. Microbial Ecology 58: 189-198

8.

Lambais MR, Crowley DE, Cury JC, Bull RC, Rodrigues RR (2006) Bacterial diversity
in tree canopies of the Atlantic forest. Science 312: 1917-1917

9.

Ley RE, Lozupone CA, Hamady M, Knight R, Gordon JI (2008) Worlds within worlds:
evolution of the vertebrate gut microbiota. Nature Reviews Microbiology 6: 776-788

10.

Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, Sargent M, Gill SR,
Nelson KE, Relman DA (2005) Diversity of the Human Intestinal Microbial Flora.
Science 308: 1635-1638

11.

Finlay BJ, Clarke KJ (1999) Ubiquitous dispersal of microbial species. nature 400: 828828

12.

Feldhaar H, Gross R (2008) Immune reactions of insects on bacterial pathogens and
mutualists. Microbes and Infection 10: 1082-1088

9

13.

Fraune S, Augustin R, Bosch TCG (2011) Embryo protection in contemporary
immunology: Why bacteria matter. Communicative & Integrative Biology 4: 369-372

14.

Grindstaff JL, Brodie ED, Ketterson ED (2003) Immune function across generations:
integrating mechanism and evolutionary process in maternal antibody transmission.
Proceedings of the Royal Society of London Series B: Biological Sciences 270: 23092319

15.

Kudo S (2000) Enzymes responsible for the bactericidal effect in extracts of vitelline and
fertilisation envelopes of rainbow trout eggs. Zygote 8: 257-265

16.

Fuda H, Hara A, Yamazaki F, Kobayashi K (1992) A peculiar immunoglobulin M (IgM)
identified in eggs of chum salmon (Oncorhynchus keta). Developmental and
Comparative Immunology 16: 415-423

17.

Morrison C, Bird C, O'Neil D, Leggiadro C, Martin-Robichaud D, Rommens M,
Waiwood K (1999) Structure of the egg envelope of the haddock, Melanogrammus
aeglefinus, and effects of microbial colonization during incubation. Canadian Journal of
Zoology 77: 890-901

18.

Palmer C, Bik EM, DiGiulio DB, Relman DA, Brown PO (2007) Development of the
Human Infant Intestinal Microbiota. PLoS Biol 5: e177

19.

Connell JH, Slatyer RO (1977) Mechanisms of succession in natural communities and
their role in community stability and organization. Am Nat 111: 1119-1144

20.

Favier CF, Vaughan EE, De Vos WM, Akkermans ADL (2002) Molecular Monitoring of
Succession of Bacterial Communities in Human Neonates. Appl Environ Microbiol 68:
219-226

21.

Fierer N, Hamady M, Lauber CL, Knight R (2008) The influence of sex, handedness, and
washing on the diversity of hand surface bacteria. Proceedings of the National Academy
of Sciences 105: 17994-17999

22.

Jackson CR, Churchill PF, Roden EE (2001) Successional Changes in Bacterial
Assemblage Structure During Epilithic Biofilm Development. Ecology 82: 555-566

23.

Lyautey E, Jackson C, Cayrou J, Rols J-L, Garabétian F (2005) Bacterial Community
Succession in Natural River Biofilm Assemblages. Microbial Ecology 50: 589-601

24.

Anderson-Glenna MJ, Bakkestuen V, Clipson NJW (2008) Spatial and temporal
variability in epilithic biofilm bacterial communities along an upland river gradient.
FEMS Microbiology Ecology 64: 407-418

25.

Besemer K, Singer G, Limberger R, Chlup A-K, Hochedlinger G, Hödl I, Baranyi C,
Battin TJ (2007) Biophysical Controls on Community Succession in Stream Biofilms.
Appl Environ Microbiol 73: 4966-4974

10

26.

Martiny AC, Jorgensen TM, Albrechtsen H-J, Arvin E, Molin S (2003) Long-Term
Succession of Structure and Diversity of a Biofilm Formed in a Model Drinking Water
Distribution System. Appl Environ Microbiol 69: 6899-6907

27.

Székely A, Sipos R, Berta B, Vajna B, Hajdú C, Márialigeti K (2009) DGGE and TRFLP Analysis of Bacterial Succession during Mushroom Compost Production and
Sequence-aided T-RFLP Profile of Mature Compost. Microbial Ecology 57: 522-533

28.

Okabe S, Odagiri M, Ito T, Satoh H (2007) Succession of Sulfur-Oxidizing Bacteria in
the Microbial Community on Corroding Concrete in Sewer Systems. Appl Environ
Microbiol 73: 971-980

29.

Fierer N, Nemergut D, Knight R, Craine JM (2010) Changes through time: integrating
microorganisms into the study of succession. Research in Microbiology 161: 635-642

30.

Peterson D, Vecsei P, Jennings C (2007) Ecology and biology of the lake sturgeon: a
synthesis of current knowledge of a threatened North American Acipenseridae. Reviews
in Fish Biology and Fisheries 17: 59-76

31.

Baker EA, Borgeson DJ (1999) Lake Sturgeon abundance and harvest in Black Lake,
Michigan, 1975-1999. N Am J Fish Manage 19: 1080-1088

32.

Smith KM, Baker EA (2005) Characteristics of spawning lake sturgeon in the Upper
Black River, Michigan. N Am J Fish Manage 25: 301-307

33.

Forsythe PS (2010) Exogenous correlates of migration, spawning, egg deposition and egg
mortality in the lake sturgeon (Acipenser fulvescens). Ph.D. Dissertation. Department of
Fisheries and Wildlife. Michigan State University. #3417681. pp191

34.

Cherr GN, Clark WH (1982) Fine Structure of the Envelope and Micropyles in the Eggs
of the White Sturgeon, Acipenser transmontanus Richardson. Development, Growth &
Differentiation 24: 341-352

35.

Lamy D, Obernosterer I, Laghdass M, Artigas LF, Breton E, Grattepanche JD, Lecuyer E,
Degros N, Lebaron P, Christaki U (2009) Temporal changes of major bacterial groups
and bacterial heterotrophic activity during a Phaeocystis globosa bloom in the eastern
English Channel. Aquatic Microbial Ecology 58: 95-107

36.

Sekiguchi H, Watanabe M, Nakahara T, Xu B, Uchiyama H (2002) Succession of
Bacterial Community Structure along the Changjiang River Determined by Denaturing
Gradient Gel Electrophoresis and Clone Library Analysis. Appled and Environmental
Microbiology 68: 5142-5150

37.

Kline KA, Fälker S, Dahlberg S, Normark S, Henriques-Normark B (2009) Bacterial
Adhesins in Host-Microbe Interactions. Cell host & microbe 5: 580-592

38.

Rendueles O, Ghigo J-M (2012) Multi-species biofilms: how to avoid unfriendly
neighbors. FEMS Microbiology Reviews 36: 972-989

11

39.

Saurabh S, Sahoo PK (2008) Lysozyme: an important defence molecule of fish innate
immune system. Aquaculture Research 39: 223-239

40.

Braun MH, Steele SL, Ekker M, Perry SF (2009) Nitrogen excretion in developing
zebrafish (Danio rerio): a role for Rh proteins and urea transporters. American Journal of
Physiology Renal Physiology 296: F994-1005

41.

Chadwick T, Wright P (1999) Nitrogen excretion and expression of urea cycle enzymes
in the atlantic cod (Gadus morhua l.): a comparison of early life stages with adults. The
Journal of Experimental Biology 202: 2653-2662

42.

Grossart H-P, Schlingloff A, Bernhard M, Simon M, Brinkhoff T (2004) Antagonistic
activity of bacteria isolated from organic aggregates of the German Wadden Sea. FEMS
Microbiology Ecology 47: 387-396

43.

Lo Giudice A, Brilli M, Bruni V, De Domenico M, Fani R, Michaud L (2007)
Bacterium–bacterium inhibitory interactions among psychrotrophic bacteria isolated from
Antarctic seawater (Terra Nova Bay, Ross Sea). FEMS Microbiology Ecology 60: 383396

44.

Mangano S, Michaud L, Caruso C, Brilli M, Bruni V, Fani R, Lo Giudice A (2009)
Antagonistic interactions between psychrotrophic cultivable bacteria isolated from
Antarctic sponges: a preliminary analysis. Research in Microbiology 160: 27-37

45.

Rypien KL, Ward JR, Azam F (2010) Antagonistic interactions among coral-associated
bacteria. Environmental Microbiology 12: 28-39

46.

Barton AD, Dutkiewicz S, Flierl G, Bragg J, Follows MJ (2010) Patterns of Diversity in
Marine Phytoplankton. Science 327: 1509-1511

47.

Besemer K, Peter H, Logue JB, Langenheder S, Lindstrom ES, Tranvik LJ, Battin TJ
(2012) Unraveling assembly of stream biofilm communities. ISME J 6: 1459-1468

48.

Auer NA (1996) Response of spawning lake sturgeons to change in hydroelectric facility
operation. Transactions of the American Fisheries Society 125: 66-77

49.

King J, Cambray JA, Dean Impson N (1998) Linked effects of dam-released floods and
water temperature on spawning of the Clanwilliam yellowfish Barbus capensis.
Hydrobiologia 384: 245-265

12

CHAPTER 2: MICROBIAL COMMUNITY ASSEMBLY AND SUCCESSION ON LAKE
STURGEON EGG SURFACES AS A FUNCTION OF SIMULATED SPAWNING
STREAM FLOW RATE

Fujimoto M, JA Crossman, KT Scribner, and TL Marsh.
Microbial Ecology. In revision as of November 7, 2012.

Abstract
We investigated microbial succession on Lake Sturgeon (Acipenser fulvescens) egg
surfaces over the course of their incubation period as a function of simulated stream flow rate.
The primary objective was to characterize the microbial community assembly during succession
and to examine how simulated stream flow rate affect the successional process. Sturgeon eggs
were reared under three flow regimes; high (0.55 m/s), low (0.18 m/s), and variable (0.35 and
0.11 m/s alternating 12 hr intervals). Eggs were collected from each flow regime at different egg
developmental stages. Microbial community DNA was extracted from the egg surface and the
communities were examined using 16S rRNA gene-based Terminal Restriction Fragment Length
Polymorphism (TRFLP) and 454 pyrosequencing. Analysis of these datasets using Principal
Component Analysis (PCA) revealed microbial communities were clustered by egg
developmental stages (early, middle, and late) regardless of flow regimes. 454 pyrosequencing
data suggested that 90-98% of the microbial communities were composed of phyla
Proteobacteria and Bacteroidetes throughout succession. -Protebacteria was more dominant in
the early stage, Bacteroidetes became more dominant in the middle stage, and -Proteobacteria

13

became dominant in the late stage. A total of 360 genera and 5826 OTUs at 97% similarity cutoff
were associated with the eggs. Midway through egg development, the egg-associated
communities of the low flow regime had a higher diversity than those communities developed
under high or variable flow regimes. The results suggest that microbial community turnover
occurred during embryogenesis, and stream flow rate could influence the microbial succession
processes on the sturgeon egg surfaces.

14

Introduction
The process of succession involves the colonization of open space and subsequent
sequential changes in community composition. This process has been studied primarily using
plant systems [1]. Microbial succession has been recently studied in conjunction with various
host animals [2-4], host plants [5], and natural [6-7] and artificial [8-10] environments. Microbes
serve as ideal targets for studying succession because the microbial succession can be effectively
controlled and observed in an experimental setting, thus allowing for novel information to be
learned about factors affecting the succession process [11].
Results from a number of studies suggest that microbial succession is a complex process
[4-6, 11]. For example, Redford and Fierer studied microbial community assembly on the newly
synthesized cottonwood leaf surface [5]. The authors found a strong temporal effect whereby
microbial community compositions on leaves from different trees sampled on the same day were
more similar to each other than those from the same tree sampled on different dates. In contrast,
Palmer and colleagues studied microbial succession in the human infant gut, and found that
inter-individual variations dominated over temporal variation [4]. Previous studies have also
shown that microbial species richness varies with time during succession. Redford and Fierer
found a positive linear relationship between microbial species richness on the cottonwood leaf
surfaces and time [5]. In contrast, Jackson and colleagues found that species richness of
microbial communities on glass slides in aquatic environments fluctuated over time, with high
diversity in the beginning due to stochastic colonization, subsequent decline in diversity due to
species sorting, followed by an increase due to maturation [6]. These differences across different
studies could be attributed to differences in the rate of dispersal, host characteristics,
environmental conditions, disturbances, and temporal scales of their studies [11].

15

Although numerous studies have demonstrated that environmental conditions are key in
explaining microbial community assembly [12-15], few studies have examined the effect of
environmental conditions on microbial succession [7, 16]. Besemer and colleagues studied the
effect of stream flow velocity on microbial succession on ceramic coupon surfaces in streams
[16]. They found that midway through the successional process, community assemblages
developed under turbulent flow were different from those developed under other flows. Lyautey
and colleagues studied the effect of environmental factors such as light and water temperature on
microbial succession on pebble surfaces in natural streams and found that both affected the
successional process [7]. This small collection of studies suggest that microbial succession is
dependent on environmental factors and further research may be needed to better understand
such effects.
In addition, despite the solid foundation of research on microbial succession on various
hosts, no studies have yet investigated microbial succession on fish eggs. Fish eggs serve as a
good model for studying microbial succession, since egg surfaces provide an open niche for
aquatic microbes. Aquatic microbes start colonizing the egg surfaces as soon as eggs are
deposited in a stream, and the microbial community subsequently develops on the egg surface
during embryogenesis. Microbial community turnovers are expected to occur as microbes on the
egg surfaces compete for egg nutrients [17], metabolites excreted by eggs change during
embryogenesis [18-19], and microbes on the egg surfaces are selected against by host innate
immunity including lysozyme secretion [20-21]. Studies about microbial succession on egg
surfaces in streams will provide some insights about the effect of live hosts on microbial
succession by providing a comparison to previous microbial succession studies performed using
natural [7] and artificial substrates [6, 16] in streams.

16

We present here a study on microbial community succession on the Lake Sturgeon
(Acipenser fulvescens) egg surface. Lake Sturgeon populations have decreased drastically over
the past 100 years due to anthropogenic activities such as overfishing and dam construction [22].
One such population is the Black Lake population, Michigan [23-24]. Spawning habitats have
been altered since the construction of Kleber dam in 1949 on the Upper Black River, which is the
sole spawning stream for the Black Lake population. Despite recent restoration efforts, natural
recruitment is limited, which is likely attributed to high egg mortality [25]. Our primary
objective was to understand how microbial community assembly changes throughout the fish
embryonic development and to acquire fundamental knowledge about microbial community
assembly on the egg surface during succession. Characterizing microbial community assembly
on the egg surface during incubation will help illuminate the potential causal relationship
between microbes and egg mortality. We were also interested in investigating how changes in
stream flow rate affected microbial community assembly during microbial succession. Flow rate
is an important environmental factor to consider in this system because construction of dams has
altered steam flow rate [26-29], and in turn potentially affected downstream ecosystems in many
ways, including by altering the interactions between microbes and fish eggs.

Methods
Experimental Design
This experiment was conducted at a Lake Sturgeon streamside rearing facility (details in
[30]) located on the Upper Black River system in Michigan during May 2007 in the midst of the
Lake Sturgeon spawning season. Incoming river water was filtered using a sand filtration system
to remove large particulate matter before being gravity-fed in a flow-through design to

17

experimental flumes. We tested the effects of three different flow regimes on microbial
community succession on the egg surface over time. A total of six flume channels were used
with two replicates for each flow regime. The first flow regime consisted of a constant high flow
velocity (0.55 ± 0.01m/s) representing a fast current section of a natural stream which embryos
experience in the natural river setting [26]. The second flow regime was a low flow velocity
(0.18 ± 0.01m/s) that represented both slower areas in the river and minimum flow that eggs
naturally experience during dry spring seasons. A third flow regime was a variable flow. This
variable flow regime was set to be high for 12 hours and low for 12 hours (0.35 ± 0.01m/s and
0.11 ± 0.01m/s, respectively), which is typical of many hydro electric dams that operate during
periods of peak electrical demand [26]. Gametes used in this study were collected from adult
Lake Sturgeon captured in the act of spawning in the Upper Black River, Michigan. Eggs were
obtained from one female and were fertilized with milt from two males that were selected
randomly from a pool of candidates. Immediately upon fertilization, approximately 100 eggs
were placed on plexi-glass plates (7.6cm x 5cm). Eggs were allowed to adhere to plates in
standing stream water. Thus, the conditions of initial colonization were standardized across the
eggs to be exposed to the three different flow regimes. Plates were then placed within
experimental flume channels constructed from a 3" PVC pipe cut lengthwise. All plates were
kept at a water depth of 2 cm across treatments. The water source throughout the experiment was
exactly the same for all six flumes so infusion of stream microbes was constant across the flumes.
Six to ten eggs were randomly collected from each flume channel at five embryonic
developmental stages (Day 2, Day 3, Day 5, Day 6, and Day 7) and were immediately preserved
in 70% ethanol. Over 90% of all eggs that hatched did so between Day 9 and Day 11. The total
number of samples collected in this study was 30 (6 flume channels * 5 developmental stages).

18

o

Water temperature ( C) was also measured at hourly intervals throughout the experiment.
Dissolved oxygen concentration of the source water was above 9 mg/L throughout the
experiment.

DNA extraction and Terminal Restriction Fragment Length Polymorphism (TRFLP)
The thirty microbial community samples were subjected to DNA extraction. Microbial
community genomic DNA was extracted from the surfaces of 4 or 5 eggs per sample using the
PowerSoil

TM

Kit (MO BIO Laboratories Inc., CA) following bead beating according to the

manufacturer’s protocol. 16S rRNA gene based TRFLP [31-32] was performed to compare
microbial community assembly across the 30 samples. TRFLP technique has been used to
compare microbial community composition across environmental gradients [12, 33-34],
locations [12, 35-36], and times [8, 10, 37]. The detailed procedures are as follows. 16S rDNA
amplification was performed using the universal bacterial primers 63F (5’-CAG GCC TAA CAC
ATG CAA GTC-3’) (5’FAM-labelled) and 1389R (5’-ACG GGC GGT GTG TAC AAG-3’)
(unlabelled) [38-39]. PCR reaction was conducted in a 100 µL reaction volume, containing 20 to
-1

40µL template DNA (1 to 4 ng µL ), 5.0 U of Taq DNA polymerase (Invitrogen Corp.,
Carlsbad, CA), in a final concentration of 0.2µM of each primer, 0.25mM of each
deoxynucleoside triphosphate, 10mM Tris-HCl, 50mM KCl, and 1.5mM MgCl2. PCR was
performed under the following cycle conditions: an initial denaturation step at 94°C for 5 min
and 30 cycles of denaturation at 94 °C for 30s, annealing at 55°C for 30s, and extension at 72°C
for 110s. A final extension step at 72 °C for 7 min was then performed. The PCR product was
purified using QIAquick PCR purification kit (Qiagen) according to the manufacturer’s protocol.

19

The purified PCR products were subjected to enzyme digestion with either HhaI or MspI (Gibco
BRL). The reaction mixture contained 2.0 µL of 10X reaction buffer (Gibco BRL), 0.3 µL of
enzyme (20U/µL, Gibco BRL), about 200ng of purified PCR product and pure water to a final
o

volume of 20 µL. The enzyme digestion was carried out for 2 hours at 37 C. Two technical
replicates (10µL each) of the digested DNA samples were sent to Michigan State University’s
sequencing facility and the DNA fragments were separated on an ABI 3100 Genetic Analyzer
automated sequencer (Applied Biosystems Instruments, Foster City, CA) in GeneScan mode.
The 5’ terminal restriction fragments (TRFs) were detected by excitation of the 6-FAM molecule
attached to the forward primer. The sizes and abundance (peak height) of the terminal fragments
were calculated using GeneScan 3.7. The resultant peak heights were filtered to eliminate peaks
with a height below the background noise threshold (set at 50 fluorescence units). Each terminal
fragment corresponds to a phylotype, and peak height indicates relative abundance of a
phylotype. In order to align TRF peaks across the 30 samples, the TRFLP profiles were
processed with T-Align (http://inismor.ucd.ie/~talign/index.html) and the output of T-Align was
used for the microbial community analysis. A total of 123 phylotypes and 130 phylotypes were
detected from the 30 samples using endonuclease HhaI and MspI, respectively. On average, 26.6
and 25.6 phylotypes were detected per sample for HhaI and MspI, respectively.

Microbial community analysis using TRFLP data
We employed the Bray-Curtis dissimilarity index [40] to compare microbial community
composition among samples from different flow treatments and egg developmental stages. A
dissimilarity index of 0 indicates that the community compositions of the two samples are
identical, whereas an estimated index of 1.0 indicates that the community compositions of the

20

two samples are completely different. We employed general linear models with the dissimilarity
index as the dependent variable to assess the effect of time and flow treatments on the index.
Principal component analysis (PCA) was performed using TRFLP data from the 30 samples to
elucidate underlying patterns across samples. The data consisted of 25 columns which
represented 25 major phylotypes that had 3% or higher relative abundance for at least one of the
30 samples, and 30 rows which represented relative abundance of each major phylotype in the 30
samples. The scores of principal component 1 (PC1) and 2 (PC2) were used to elucidate the
temporal similarity of the microbial community composition, and the loadings of PC1 and PC2
were used to elucidate the distribution patterns of the 25 major phylotypes. Phylotype richness S
(S = the number of distinct terminal restriction fragments in each sample) was determined from
the TRFLP profiles. All distinct TRFs, including both major and minor phylotypes, were
included in this analysis. Prior to analyzing the relationship between phylotype richness and time,
extremely high or low total fluorescence signals caused by over or under loading of digested
DNA samples were removed to eliminate non-biological effects on phylotype richness. A
general linear model was used to investigate the relationship between the microbial phylotype
richness and time. A quadratic term for time was also included in the model. We conducted all
analyses using both HhaI and MspI, but both showed similar results (data not shown); therefore
we present only the HhaI results. The general linear model and PCA were conducted using R
version 2.10.0 [41]. The Bray-Curtis dissimilarity matrix was generated in R using the “ecodist”
package.

454 Pyrosequencing

21

To characterize the microbial community at different time points and under different flow
regimes, nine samples representing three flow regimes (High, Low and Variable) at three time
points (Day 2, Day 5, and Day 7) were assigned for pyrosequencing. Since the TRFLP data
suggested that variation among two replicated flows within the same flow regime at the same
time point was small, the extracted genomic DNAs of the two replicated flows within the same
flow regime were pooled. To evaluate the reproducibility of samples, technical replicates were
included for the three Day 5 samples (High, Low and Variable at Day 5). Hypervariable region
V3 - V5 in 16S rRNA gene was amplified using a forward primer 357F (5’CCTACGGGAGGCAGCAG-3’) and a reverse primer 926R (5’-CCGTCAATTCMTTTRAGT3’) as previously described [42-44]. 454 ‘A’ adapter and tag sequences were contained in the
reverse primer, and ‘B’ adapter was contained in the forward primer. PCR amplification was
performed in 75µL reaction, using 3U of AccuPrime Taq HiFi (Invitrogen, Grand Island, NY),
7.5µL of supplied 10X buffer II, 1.5µL of 10µM primers, and approximately 30ng of template
DNA measured by Nanodrop ND1000 (Thermo Scientific Inc). The PCR cycle condition was as
o

o

follows: denaturation at 95 C for 2 minutes followed by 30 cycles of denaturation at 95 C for 20
o

o

seconds, annealing at 50 C for 30 seconds, extension at 72 C for 5 minutes [44]. The PCR
amplicons were cleaned using Agencourt AmPure XP Beads (Beckman Coulter, Inc., Brea, CA)
and pyrosequencing of the amplicons was performed using 454 GS FLX titanium platform (454
Life Science, Branford, CT) at the research technology support facility at Michigan State
University.

Pyrosequencing data processing and analysis

22

Raw sequence reads were processed using Ribosomal Database Project (RDP) pipeline
[45] to sort the data by tag sequence, to trim tag and primer sequences, and to filter out low
quality sequences with minimum quality score of 20 (probability threshold of 0.01) and
minimum read length of 300bp. The taxonomy of the filtered reads was assigned using RDP
Classifier at a bootstrap threshold of 80% [46]. Bray-Curtis dissimilarity among the 9 samples
was determined using the classifier output at the genus level. OTUs of the filtered reads at 97%
similarity cutoff were determined using RDP complete-linkage clustering algorithms with 0.03
maximum cluster distances. The underlining patterns of the 9 samples were analyzed by PCA
using OTUs at the 97% similarity. Jaccard index tree of OTUs at the 97% similarity level was
constructed using RDP pipeline. The diversity index including species evenness (E) for each
sample was determined by RDP pipeline diversity index estimator using the 97% OTUs.
Rarefaction curves of the 97% OTUs for the samples were generated using RDP pipeline. To
assess the relationship between the number of OTUs and time, the 97% OTUs were rarefied at
6000 reads per sample.

Results
Analysis on microbial community succession using TRFLP
Using TRFLP analysis, we detected considerable temporal variation in microbial
community composition on Lake Sturgeon eggs across sequential embryonic developmental
stages (Figure 2.1). The temporal variation in microbial community composition was
consistently observed in all six flume channels (two replicates for each of the 3 flow regimes).
Microbial community compositions estimated from samples collected from the same
developmental stage were more similar to each other than those from different developmental

23

Figure 2.1. Temporal and flow rate effects on microbial community composition during
microbial succession.
25 major phylotypes that had 3% or higher relative abundance for at least one sample were
included. The rest were grouped into “others”. Each color/pattern represents a unique terminal
fragment (phylotype, PT). The experiment consists of 3 flow regimes (High: Flows 1 and 2, Low:
Flows 3 and 4, and Variable: Flows 5 and 6) sampled at day 2, 3, 5, 6, and 7. HhaI digested
TRFLP data were used for the analysis. For interpretation of the references to color in this and all
other figures, the reader is referred to the electronic version of this dissertation.
24

stages, regardless of flow regimes (Table 2.1). For example, the mean Bray-Curtis dissimilarities
within Day 2 samples and between Day 2 and Day 3 samples were 0.301 and 0.594, respectively
(Table 2.1). These two means were statistically significantly different (F1,49 = 135, p < 0.001).
The dissimilarity between microbial community compositions increased during the course of
embryogenesis as the community departed from the initial community (Table 2.1, left column).

Table 2.1. Bray-Curtis dissimilarity index matrix using TRFLP data summarized by day
Day 2

Day 3

Day 5

a

Mean
Mean
Mean
(sd)
(sd)
(sd)
0.301
Day 2
(0.083)
0.594
0.426
Day 3
(0.082)
(0.082)
0.624
0.443
0.444
Day 5
(0.073)
(0.114)
(0.104)
0.639
0.510
0.497
Day 6
(0.068)
(0.103)
(0.141)
0.747
0.592
0.585
Day 7
(0.064)
(0.092)
(0.110)
a. Mean Bray-Curtis dissimilarity across all flow regimes.

Day 6

Day 7

Mean
(sd)

Mean
(sd)

0.490
(0.134)
0.518
(0.140)

0.369
(0.095)

Stream flow rate had a more subtle effect on the composition of the microbial
communities than egg developmental stage (Figure 2.1). Microbial communities were similar at
the early (Day 2) and late (Day 7) developmental stages, regardless of flow regimes. However,
the microbial composition of the low flow regime departed from that in the high and variable
flow regimes during mid-developmental stages. This is evident in the diagonal array of the
dissimilarity index in Table 2.1 that shows an increasing variance of dissimilarity (which we
attribute to differences in flow). At Day 5, the microbial community composition differed

25

between high (flow 1 and 2) and low (flow 3 and 4) flow regimes (Figure 2.1: Day 5). The mean
dissimilarity within the same flow regime (high-high or low-low) was 0.368, whereas the mean
dissimilarity between high and low flow regimes was 0.478. Microbial communities reared in the
variable flow regime were more similar to that in the high flow regime (dissimilarity = 0.380)
than to the low flow regime (dissimilarity = 0.554), suggesting that the high flow portion of the
variable flow regime had a more dominant effect on microbial community structure than the low
flow.
Principal component plots also showed the strong temporal clustering of microbial
communities on the egg surfaces (Figure 2.2). Microbial communities sampled from the same or
close developmental stages were clustered together regardless of flow regime. Microbial
communities were clustered into three different embryonic developmental stages; the early (Day
2), the middle (Day 3, Day 5, and Day 6) and the late (Day 7), respectively. This PCA plot shows
that the temporal effect was more dominant in explaining microbial community assembly than
flow rate effects. We also found that certain microbial phylotypes were more strongly associated
with certain egg developmental stages (Appendix Figure A.2.1). Out of the 25 dominant
microbial phylotypes we detected using HhaI, we found that 8 phylotypes were predominantly
associated with the early stage, 7 phylotypes peaked in the middle, and 9 phylotypes were
preferentially associated with the late stages of egg development. This trend was also depicted in
a loading principal component plot (Figure 2.3).
There was a significant positive linear relationship between microbial phylotype richness
on egg surfaces and time (Figure 2.4 solid line: F1,25 = 4.73, p = 0.039). In contrast, there was no
significant main effect of flow regime on microbial phylotype richness according to a linear
regression model (F2,24 = 0.658, p = 0.53). However, the relationship between microbial

26

Figure 2.2. Principal component score plots revealing the temporal clustering of microbial
community assemblages on egg surfaces.
PC1 and PC2 account for 27.9% and 13.6% of the data variations, respectively. HhaI digested
TRFLP data were used for the analysis.

27

Figure 2.3. Principal component loading plots displaying temporal distributions of 25
major microbial phylotypes (PT) associated with the egg surface.
PC1 and PC2 account for 27.9% and 13.6% of the data variations, respectively. HhaI digested
TRFLP data were used for the analysis.

28

Figure 2.4. Microbial phylotype richness and time relationships during succession across
flow regimes (High ○, Low +, and Variable ∆).
2

The straight solid line is a linear regression line for all data points (F1,25 = 4.73, p = 0.039, R =
0.16). The dotted line is a positive quadratic regression for the low flow regime (F2,5 = 10.26, p
2

= 0.017, R = 0.80). The breaking line is a negative quadratic regression for the high flow regime
2

(F2,7 = 1.543, p = 0.28, R = 0.31). The dotted-dash line is a negative quadratic regression for
2

variable flow regime (F2,6 = 13.1, p = 0.006, R = 0.81). The phylotype richness determined by
HhaI digested TRFLP was used for this analysis.

29

phylotype richness and time was dependent on flow regime. At the low flow regime, the
relationship between microbial phylotype richness and time was negative quadratic (Figure 2.4
dotted line: F2,5 = 10.26, p = 0.017) with the highest richness at the middle developmental stage.
In contrast, under the high and variable flow regimes, there was a moderate positive quadratic
relationship between microbial phylotype richness and time (Figure 2.4 breaking line (high): F2,7
= 1.543, p = 0.28; dotted-dash line (variable): F2,6 = 13.1, p = 0.006) with lower phylotype
richness at the middle developmental stage. In other words, at Day 5, eggs reared in the low
flow regime were associated with a relatively large number of microbial phylotypes, while eggs
reared in high or variable flow regimes were associated with a fewer number of phylotypes.

Microbial community analysis using 454 pyrosequencing
To characterize the microbial community assembly at different egg developmental stages
and under different flow regimes, nine samples representing three flow regimes (High, Low and
Variable) at three time points (Day 2, Day 5, and Day 7) were subjected to pyrosequencing. We
obtained an average number of 13468 ± 7779 reads with an average length of 450.08 ± 1.97bp
per sample. These reads were classified using RDP classifier into 23 phyla, 45 classes, and 360
genera across the 9 samples. The average numbers of phyla, classes, and genera per sample were
14.5 ± 2.6, 25.3 ± 4.5, and 138.7 ± 20.2, respectively. Two technical replicates for Day 5 samples
showed an almost identical community composition at each taxonomic level, indicating that the
pyrosequencing data were reproducible.
A temporal compositional trend during microbial succession was detected at the
phylum/class level (Figure 2.5). Two phyla, Proteobacteria and Bacteroidetes, comprised 90 to

30

98% of the egg surface microbial community throughout egg development. Proteobacteria was
more dominant in the early (58 to 74%) and late (71 to 82%) egg developmental stages, while
Bacteroidetes was dominant in the middle stage (63% for high and 81% for variable flow regime)
except under the low flow regime (36%). The dominant classes of the phylum Proteobacteria
included , , and -Proteobacteria. -Proteobacteria was more dominant in the early egg
developmental stage and -Proteobacteria became more dominant by the late stage.
The temporal trend was also found at the genus level of analysis (Figure 2.6).
Flavobacterium was one of the most dominant genera in the egg surface microbial community
during succession and showed a strong temporal trend (Figure 2.6). Genus Flavobacterium
accounted for a large proportion of the egg surface microbial community in the early (20 to 40%),
middle (25 to 80%), and late (9 to 13%) egg developmental stages. The large variation in

Figure 2.5. Characterization of the egg surface microbial community assembly using the
RDP classifier output at the phylum level (and class level for Proteobacteria).
31

Figure 2.6. Characterization of the egg surface microbial community assembly using the
RDP classifier output at the genus level.

Table 2.2. Bray-Curtis dissimilarity matrix using the RDP classifier output at genus level
High
Low
Var
High
Low
D2
D2
D2
D5
D5
Low D2 0.199
Var D2 0.261
0.226
High D5 0.366
0.511
0.561
Low D5 0.384
0.465
0.512
0.352
Var D5 0.429
0.581
0.645
0.185
0.481
High D7 0.684
0.633
0.652
0.688
0.513
Low D7 0.754
0.731
0.750
0.740
0.601
Var D7 0.642
0.597
0.606
0.618
0.442
*Unclassified genera were not included in this analysis.

32

Var
D5

High
D7

Low
D7

0.772
0.805
0.712

0.323
0.285

0.293

Flavobacterium relative abundance during the middle stage can be attributed to flow treatment.
At Day 5, Flavobacterium accounted for 55 to 80% of the egg surface microbial community
under high and variable flow regimes, but only 25% under the low flow regime. Other genera
also exhibited a temporal trend in abundance during succession. Genera such as Brevundimonas,
Undibacterium, Massilia, Acidovorax, and Rheinheimera decreased as egg development
progressed, Flectobacillus and Fluviicola peaked at intermediate periods, and Rhodobacter,
Catellibacterium, and Devosia (all -Proteobacteria) were dominant towards the end of
development. Although almost all sequences were classified at phylum/class level, 12 to 47% of
the egg surface microbial communities per sample were unclassified at genus level, and the
proportion of the unclassified genera increased as egg development progressed. We characterized
a total of 360 genera in this analysis, but only two, Flavobacterium and Albidiferax, maintained a
relative abundance over 2% throughout the entire incubation period.
Bray-Curtis dissimilarity analysis conducted at the genus level showed that microbial
communities taken from the same time points were more similar to each other than to those from
other time points (Table 2.2). The dissimilarity increased during the course of embryogenesis as
the later communities departed from the initial community (Table 2.2, left column). At Day 5,
the microbial community under low flow regime differed from those under high and variable
flow, similar to the results seen with TRFLP.
RDP complete cluster linkage algorithms revealed the existence of 5826 distinct OTUs
across the samples at 97% similarity cutoff with the average OTUs 1216.8 ± 201.0 per sample.
Rarefaction curves of OTUs at the 97% similarity for the 9 samples including technical replicates
are shown in Appendix Figure A.2.2. To assess the underlining patterns across different time
points and flow regimes, the data matrix of the 97% OTUs was subjected to PCA. The first two

33

principal components captured the temporal trend in microbial community structures, while PC3
accounted for differences between low flow regime and both high and variable flow regimes at
Day 5 (Figure 2.7). The Jaccard distance matrices using the 97% OTUs also grouped samples
collected at the same time point together (Appendix Figure A.2.3), while separating low flow
regime from high and variable flow regimes within the Day 5 group. The 97% OTUs rarefied at
6000 reads and time relationship showed a similar trend to what was found using TRFLP (Figure
2.8). The Shannon species evenness (E) calculated using the 97% OTUs revealed that microbial
communities developed at Day 5 under high and variable flow were unevenly distributed (Table
2.3), which is congruent with the dominance of genus Flavobacterium in these communities.

34

Figure 2.7. Principal component score plots using OTUs defined at 97% similarity cutoff.
PC1, PC2, and PC3 accounted for 15.3%, 12.2%, and 10.8% of the data variations, respectively.
H, L, V, in the figure denotes High, Low, and Variable flow, respectively. Two technical
replicates of Day 5 samples were included.

35

Figure 2.8. Microbial phylotype richness (97% OTUs rarefied at 6000 reads) and time
relationships during succession across flow regimes.

Table 2.3. Estimated species evenness (E) using OTUs at 97% similarity cutoff
Flow Regime
High
Low
Variable

Day 2
0.635
0.684
0.721

Day 5
0.616
0.787
0.501

Day 7
0.754
0.731
0.794

36

Discussion
To our knowledge, this is the first study that documents microbial succession on fish egg
surfaces using next generation sequencing. Although we obtained the average of 13,400 reads
per sample using the pyrosequencing, our rarefaction curves indicate that microbial community
assembly on the fish egg surfaces were not completely covered. However, our integrative
approach of combining analysis of microbial community structures using TRFLP and subsequent
characterization of microbial community assembly using pyrosequencing was effective in
comprehensively identifying microbial community succession patterns. Both our TRFLP and
pyrosequencing results clearly demonstrated that microbial species replacement was occurring
on fish egg surfaces. The fish egg microbial communities were clustered into three different
embryonic developmental stages: the early, the middle, and the late, which is similar to what
Besemer and colleagues found with microbial succession on the ceramic coupon in streams [16],
although they observed this trend over a longer time frame. The fact that we detected a temporal
trend during microbial succession at the phylum/class level suggests that the change in microbial
community assembly is not subtle, but rather drastic. This is particularly evident from our
assessment that only two genera out of 360 detected genera maintained a relative abundance of
over 2% throughout the entire incubation period.
One possible factor that may have contributed to the observed microbial community
shifts during embryogenesis is the change in the chemistry of the egg surface. Previous studies
suggest that chemistry on fish egg surfaces change during egg development [18-21]. Chadwick
and Wright studied the nitrogen excretion patterns from Atlantic Cod eggs and found that the
excretion of ammonia increased linearly while urea excretion decreased as the eggs developed
[18]. Although nitrogen excretion patterns have never been studied in sturgeon, this type of

37

metabolite excretion could contribute to a change in the microbial community on the egg surface.
Fish eggs are also known to secrete antimicrobial substances such as lysozyme [20-21] thus
bactericidal activities might select for or against certain microbial species. One study on a
relationship between egg innate immunity of small freshwater metazoan and egg microbial
communities demonstrated that microbial communities on egg surfaces were altered when the
type of antimicrobial peptides on eggs changed from maternally provisioned to that secreted by
eggs [47-48]. Because over 90% of the egg surface microbial communities we found were
composed of gram negative phyla Proteobacteria and Bacteroidetes, it is possible that microbial
community assembly on the egg surface was shaped by egg-excreted lysozyme, which acts on
peptidoglycan layers of gram positive bacteria [49].
The observed microbial community turnover also could be attributed to changes in the
microbial community composition of the surrounding water during the experiment. The
microbial community in freshwater streams varies temporally and spatially [37, 50-51]. We may
have observed a microbial community that stochastically dispersed onto the egg surface from the
water column [52-53]. However, the structure of the egg-associated microbial community
characterized in this study differed from the structure of the typical aquatic microbial community
described previously [54]. Despite having a similar high prevalence of β, α, and -Proteobacteria
and Bacteroidetes as commonly seen in freshwater communities, our observed egg surface
microbial community assemblages had almost undetected levels of the phylum Actinobacteria, a
microbe typically dominant in freshwater environments. In fact, we collected water samples from
our spawning stream during the spawning season of a different year (Fujimoto et al. unpublished
study) and also detected significant numbers of the phylum Actinobacteria in the water column
using 454 pyrosequencing, but greatly diminished numbers on the egg surfaces reared in the

38

stream water. In addition, the dominance of over 50% relative abundance of Flavobacterium in
the high and variable flow regimes at Day 5 deviated strongly from freshwater microbial
communities. These findings are revealing in suggesting that dispersal from water column may
not be a strong factor in shaping egg surface microbial community assembly.
The microbial community turnover we observed on fish egg surfaces occurred during a 7
day period, which is significantly shorter than the time frame in which microbial succession is
typically observed on non-living substrate surfaces [6-7, 16, 50]. Our results suggest that
microbial species sorting on the egg surface can occur in a short time frame, perhaps due to the
living egg-related effects such as host innate immunity [21], secretion of metabolites [18], and
provision of three dimensional structures, unlike the comparably long time required for such
sorting to be observed on abiotic substrate surfaces where species sorting after initial
colonization is limited by resource competition and microbe-microbe interaction [50].
Another finding of our study was the effect of flow rate on microbial community
assembly. Despite the lack of flow effect on the early and late egg surface microbial
communities, the differences in phylotype diversity observed between eggs reared under low
versus high and variable flow regimes at Day 5 were revealing. These findings suggest that the
microbial community compositions were uniform at the early stage, diverged toward the middle
stage dependent on the flow rate experienced by the eggs, and converged at the late stage of
embryogenesis. A similar trend was found by Besemer and colleagues [16] who observed a
significant effect of the flow velocity on ceramic surface microbial communities midway through
the succession process but not in initial and mature microbial communities.
One possible explanation for the flow effect could be that high flow rate selected for
certain microbial species that have the ability to adhere to the egg surface under greater shear

39

force. One previous study tested the effect of flow velocity on microbial community diversity in
water pipes found that increasing flow velocity in pipes lowered microbial community diversity
[55]. At low flow rate, the selection pressure was presumably relaxed; therefore, many different
microbial species with lower adhesive capabilities could stay on the egg surface. The dominance
of Flavobacterium in the high and variable flow regimes could be related to the fact that it is a
particularly adhesive microbe [56-57]. Another possibility is that increased flow velocity
increased the concentration of dissolved oxygen, thereby resulting in selection for aerobes that
have a high tolerance for oxygen. However, we do not know why the effect of flow on
community assembly was observed in the middle of the embryogenesis, but not observed at the
early and late time points.
It is important to consider the potential implications of the effect of flow rate on microbe
– egg interactions for host life history. Another study by our group [25] showed that over 80%
of egg mortality occurs during the first half of the incubation period, which is from Day 0 to Day
5 for our experimental conditions. Flavobacterium, the dominant genus in both high and variable
flow at Day 5, encompasses ranges of species including known fish pathogens such as F.
columnare [58], F. psychrophilum [59], and F. branchiophilum [60]. On the other hand, under
low flow regime at Day 5, phylum Proteobacteria, which also includes some fish pathogens such
as genus Aeromonas [61], was dominant. The net effect of flow velocity on microbe- host
interaction is difficult to determine. However, it is important to note that fish eggs deposited in
natural streams may not receive the extremely high flow velocity used in this experiment, since
the flow at the bottom of streams is lower than that in middle of water column due to the effect
of the boundary layer [62].

40

In this study, we focused on flow rate as a key environmental covariate that could
influence microbe – host interactions. However, water temperature is also an important
environmental covariate that may affect the microbial community assembly in aquatic systems [7,
63]. In this experiment, all eggs reared in the 6 flume channels were exposed to water from the
same source and thus experienced the same water temperature throughout the experiment.
However, the daily mean ambient water temperature gradually increased during the course of the
experiment (Appendix Figure A.2.4), which may account for the temporal variation of microbial
community composition that we observed during embryogenesis.
One implication of this study is that natural stream systems could create variation in
microbial community succession due to differences in natural flow rate observed in streams
during the spawning season. Streams may have significantly less precipitation during one
spawning season than other average seasons simply by chance, which could lead to significantly
lower stream flow rates during spawning events. Our analysis suggests that we should expect
more microbial phylotypes adhering to egg surfaces during the critical period of embryogenesis
in such dry spawning seasons.
The variable flow regime, which mimicked dam-manipulated flow, resulted in similar
successional patterns to what we observed under the high flow regime. Although the high and
low flows were alternated in the variable flow regime, the high flow operation period of the
variable flow might have had a dominant effect on microbial community composition. High flow
rate of the simulated stream could have acted as a disturbance to microbial community formation
on the egg surface. Since naturally occurring eggs potentially experience high flow during
embryogenesis, our result implies that dam operations that include periods of high flow would
not alter the natural course of successional patterns on egg surfaces. However, not all dams are

41

operated in the same way as this experiment (alternating 12 hour intervals), thus we cannot
generalize our findings to all dam systems. Furthermore, a dam could potentially alter water
chemistry, temperature, nutrients, and water microbial community composition in a spawning
stream [64], which in turn could influence the microbial communities during succession.
In conclusion, this is the first microbial succession study conducted on fish eggs.
Although previous studies have documented succession patterns on abiotic substrates in streams,
our study breaks new ground in demonstrating that microbial community shifts can occur in a
remarkably short period on a biotic surface. We hope this study will trigger more studies in this
area and encourage other researchers to delve deeper into understanding the underlying
mechanisms behind the patterns we observed.

42

Appendix

43

Appendix Figure A.2.1. Examples of the association of certain microbial phylotypes with
certain egg developmental stages.
2

The dotted line (linear coefficient = -1.89, R = 0.44), solid line (quadratic coefficient = -2.83,
2

2

R = 0.23), and breaking line (linear coefficient = 1.61, R = 0.55) represent regression lines for
PT3, PT7, and PT16, respectively. HhaI data were used for the analysis.

44

Appendix Figure A.2.2. Rarefaction analysis for pyrosequencing data.

45

Appendix Figure A.2.3. Jaccard index tree constructed using OTUs at 97% similarity
cutoff.

46

Appendix Figure A.2.4. Ambient water temperature throughout the egg incubation periods.
Solid line is a linear regression line for daily mean temperature.

47

References

48

References

1.

Connell JH, Slatyer RO (1977) Mechanisms of succession in natural communities and
their role in community stability and organization. Am Nat 111: 1119-1144

2.

Favier CF, Vaughan EE, De Vos WM, Akkermans ADL (2002) Molecular monitoring of
succession of bacterial communities in human neonates. Applied and Environmental
Microbiology 68: 219-226

3.

Fierer N, Hamady M, Lauber CL, Knight R (2008) The influence of sex, handedness, and
washing on the diversity of hand surface bacteria. Proceedings of the National Academy
of Sciences of the United States of America 105: 17994-17999

4.

Palmer C, Bik EM, DiGiulio DB, Relman DA, Brown PO (2007) Development of the
human infant intestinal microbiota. PLoS Biology 5: e177

5.

Redford A, Fierer N (2009) Bacterial succession on the leaf surface: A novel system for
studying successional dynamics. Microbial Ecology 58: 189-198

6.

Jackson CR, Churchill PF, Roden EE (2001) Successional changes in bacterial
assemblage structure during epilithic biofilm development. Ecology 82: 555-566

7.

Lyautey E, Jackson C, Cayrou J, Rols JL, Garabétian F (2005) Bacterial community
succession in natural river biofilm assemblages. Microb Ecol 50: 589-601

8.

Martiny AC, Jorgensen TM, Albrechtsen H-J, Arvin E, Molin S (2003) Long-term
succession of structure and diversity of a biofilm formed in a model drinking water
distribution system. Applied and Environmental Microbiology 69: 6899-6907

9.

Okabe S, Odagiri M, Ito T, Satoh H (2007) Succession of sulfur-oxidizing bacteria in the
microbial community on corroding concrete in sewer systems. Applied and
Environmental Microbiology 73: 971-980

10.

Székely A, Sipos R, Berta B, Vajna B, Hajdú C, Márialigeti K (2009) DGGE and TRFLP analysis of bacterial succession during mushroom compost production and
sequence-aided T-RFLP profile of mature compost. Microb Ecol 57: 522-533

11.

Fierer N, Nemergut D, Knight R, Craine JM (2010) Changes through time: integrating
microorganisms into the study of succession. Res Microbiol 161: 635-642

12.

Fierer N, Jackson RB (2006) The diversity and biogeography of soil bacterial
communities. Proc Natl Acad Sci USA 103: 626-631

49

13.

Harris RN, Brucker RM, Walke JB, Becker MH, Schwantes CR, Flaherty DC, Lam BA,
Woodhams DC, Briggs CJ, Vredenburg VT, Minbiole KPC (2009) Skin microbes on
frogs prevent morbidity and mortality caused by a lethal skin fungus. ISME J 3: 818-824

14.

Dethlefsen L, McFall-Ngai M, Relman DA (2007) An ecological and evolutionary
perspective on human–microbe mutualism and disease. Nature 449: 811-818

15.

Besemer K, Singer G, Hodl I, Battin TJ (2009) Bacterial community composition of
stream biofilms in spatially variable-flow environments. Appl Environ Microbiol 75:
7189-7195

16.

Besemer K, Singer G, Limberger R, Chlup A-K, Hochedlinger G, Hödl I, Baranyi C,
Battin TJ (2007) Biophysical Controls on Community Succession in Stream Biofilms.
Appl Environ Microbiol 73: 4966-4974

17.

Al-Holy MA, Rasco BA (2006) Characterization of salmon (Oncorhynchus keta) and
sturgeon (Acipenser transmontanus) caviar proteins. J Food Biochem 30: 422-428

18.

Chadwick T, Wright P (1999) Nitrogen excretion and expression of urea cycle enzymes
in the atlantic cod (Gadus morhua l.): a comparison of early life stages with adults. The
Journal of Experimental Biology 202: 2653-2662

19.

Steele SL, Chadwick TD, Wright PA (2001) Ammonia detoxification and localization of
urea cycle enzyme activity in embryos of the rainbow trout (Oncorhynchus mykiss) in
relation to early tolerance to high environmental ammonia levels. J Exp Biol 204: 21452154

20.

Kudo S (2000) Enzymes responsible for the bactericidal effect in extracts of vitelline and
fertilisation envelopes of rainbow trout eggs. Zygote 8: 257-265

21.

Saurabh S, Sahoo PK (2008) Lysozyme: an important defence molecule of fish innate
immune system. Aquacult Res 39: 223-239

22.

Adams PB, Grimes C, Hightower JE, Lindley ST, Moser ML, Parsley MJ (2007)
Population status of North American green sturgeon, Acipenser medirostris. Environ Biol
Fish 79: 339-356

23.

Baker EA, Borgeson DJ (1999) Lake Sturgeon abundance and harvest in Black Lake,
Michigan, 1975-1999. N Am J Fish Manage 19: 1080-1088

24.

Smith KM, Baker EA (2005) Characteristics of spawning lake sturgeon in the Upper
Black River, Michigan. N Am J Fish Manage 25: 301-307

25.

Forsythe PS (2010) Exogenous correlates of migration, spawning, egg deposition and egg
mortality in the lake sturgeon (Acipenser fulvescens). Ph.D. Dissertation. Department of
Fisheries and Wildlife. Michigan State University. #3417681. pp191

50

26.

Auer NA (1996) Response of spawning lake sturgeons to change in hydroelectric facility
operation. T Am Fish Soc 125: 66-77

27.

Haxton TJ (2006) Characteristics of a lake sturgeon spawning population sampled a half
century apart. J Great Lakes Res 32: 124-130

28.

Jager HI, Chandler JA, Lepla KB, Van Winkle W (2001) A theoretical study of river
fragmentation by dams and its effects on white sturgeon populations. Environ Biol Fish
60: 347-361

29.

Paragamian VL, Kruse G, Wakkinen V (2001) Spawning habitat of Kootenai River white
sturgeon, post-Libby Dam. N Am J Fish Manage 21: 22-33

30.

Crossman JA (2008) Evaluating collection, rearing, and stocking methods for lake
sturgeon (Acipenser fulvescens) restoration programs in the great lakes. Ph.D.
Dissertation. Department of Fisheries and Wildlife. Michigan State University. #3331889.
pp.192

31.

Liu W, Marsh T, Cheng H, Forney L (1997) Characterization of microbial diversity by
determining terminal restriction fragment length polymorphisms of genes encoding 16S
rRNA. Applied and Environmental Microbiology 63: 4516-4522

32.

Marsh TL (1999) Terminal restriction fragment length polymorphism (T-RFLP): An
emerging method for characterizing diversity among homologous populations of
amplification products. Current Opinion in Microbiology 2: 323-327

33.

Konstantinidis KT, Isaacs N, Fett J, Simpson S, Long DT, Marsh TL (2003) Microbial
diversity and resistance to copper in metal-contaminated lake sediment. Microb Ecol 45:
191-202

34.

Tom-Petersen A, Leser TD, Marsh TL, Nybroe O (2003) Effects of copper amendment
on the bacterial community in agricultural soil analyzed by the T-RFLP technique. FEMS
Microbiol Ecol 46: 53-62

35.

Dunbar J, Ticknor LO, Kuske CR (2001) Phylogenetic specificity and reproducibility and
new method for analysis of Terminal Restriction Fragment profiles of 16S rRNA genes
from bacterial communities. Appl Environ Microbiol 67: 190-197

36.

Hackl E, Zechmeister-Boltenstern S, Bodrossy L, Sessitsch A (2004) Comparison of
diversities and compositions of bacterial populations inhabiting natural forest soils. Appl
Environ Microbiol 70: 5057-5065

37.

Anderson-Glenna MJ, Bakkestuen V, Clipson NJW (2008) Spatial and temporal
variability in epilithic biofilm bacterial communities along an upland river gradient.
FEMS Microbiology Ecology 64: 407-418

51

38.

Aburto A, Fahy A, Coulon F, Lethbridge G, Timmis KN, Ball AS, McGenity TJ (2009)
Mixed aerobic and anaerobic microbial communities in benzene-contaminated
groundwater. J Appl Microbiol 106: 317-328

39.

Quaak FCA, Kuiper I (2011) Statistical data analysis of bacterial t-RFLP profiles in
forensic soil comparisons. Forensic Sci Int 210: 96-101

40.

Bray JR, Curtis JT (1957) An ordination of the upland forest communities of Southern
Wisconsin. Ecol Monogr 27: 325-349

41.

R Development Core Team (2009) R: A launguage and environment for statistical
computing. R foundation for Statistical Computing, Vienna, Austria.

42.

Haas BJ, Gevers D, Earl AM, Feldgarden M, Ward DV, Giannoukos G, Ciulla D, Tabbaa
D, Highlander SK, Sodergren E, Methé B, DeSantis TZ, Consortium THM, Petrosino JF,
Knight R, Birren BW (2011) Chimeric 16S rRNA sequence formation and detection in
Sanger and 454-pyrosequenced PCR amplicons. Genome Res 21: 494-504

43.

Brinkman BM, Hildebrand F, Kubica M, Goosens D, Del Favero J, Declercq W, Raes J,
Vandenabeele P (2011) Caspase deficiency alters the murine gut microbiome. Cell Death
and Dis 2: e220

44.

Preidis GA, Saulnier DM, Blutt SE, Mistretta T-A, Riehle KP, Major AM, Venable SF,
Finegold MJ, Petrosino JF, Conner ME, Versalovic J (2012) Probiotics stimulate
enterocyte migration and microbial diversity in the neonatal mouse intestine. FASEB J 26:
1960-1969

45.

Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS,
McGarrell DM, Marsh T, Garrity GM, Tiedje JM (2009) The Ribosomal Database
Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res 37:
D141-D145

46.

Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naïve Bayesian Classifier for Rapid
Assignment of rRNA Sequences into the New Bacterial Taxonomy. Appl Environ
Microbiol 73: 5261-5267

47.

Fraune S, Augustin R, Anton-Erxleben F, Wittlieb J, Gelhaus C, Klimovich VB,
Samoilovich MP, Bosch TCG (2010) In an early branching metazoan, bacterial
colonization of the embryo is controlled by maternal antimicrobial peptides. Proceedings
of the National Academy of Sciences of the United States of America 107: 18067-18072

48.

Fraune S, Augustin R, Bosch TCG (2011) Embryo protection in contemporary
immunology: Why bacteria matter. Communicative and Integrative Biology 4: 369-372

49.

Callewaert L, Michiels C (2010) Lysozymes in the animal kingdom. J Biosci 35: 127-160

50.

Besemer K, Peter H, Logue JB, Langenheder S, Lindstrom ES, Tranvik LJ, Battin TJ
(2012) Unraveling assembly of stream biofilm communities. ISME J 6: 1459-1468

52

51.

Lamy D, Obernosterer I, Laghdass M, Artigas LF, Breton E, Grattepanche JD, Lecuyer E,
Degros N, Lebaron P, Christaki U (2009) Temporal changes of major bacterial groups
and bacterial heterotrophic activity during a Phaeocystis globosa bloom in the eastern
English Channel. Aquatic Microbial Ecology 58: 95-107

52.

Finlay BJ, Clarke KJ (1999) Ubiquitous dispersal of microbial species. Nature 400: 828828

53.

Dumbrell AJ, Nelson M, Helgason T, Dytham C, Fitter AH (2009) Relative roles of niche
and neutral processes in structuring a soil microbial community. ISME J 4: 337-345

54.

Newton RJ, Jones SE, Eiler A, McMahon KD, Bertilsson S (2011) A Guide to the
Natural History of Freshwater Lake Bacteria. Microbiol Mol Biol R 75: 14-49

55.

Rickard AH, McBain AJ, Stead AT, Gilbert P (2004) Shear Rate Moderates Community
Diversity in Freshwater Biofilms. Appl Environ Microbiol 70: 7426-7435

56.

Møller JD, Larsen JL, Madsen L, Dalsgaard I (2003) Involvement of a Sialic AcidBinding Lectin with Hemagglutination and Hydrophobicity of Flavobacterium
psychrophilum. Appl Environ Microbiol 69: 5275-5280

57.

Basson A, Flemming LA, Chenia HY (2008) Evaluation of adherence, hydrophobicity,
aggregation, and biofilm development of Flavobacterium johnsoniae-like isolates. Microb
Ecol 55: 1-14

58.

Schneck JL, Caslake LF (2006) Genetic diversity of Flavobacterium columnare isolated
from fish collected from warm and cold water. J Fish Dis 29: 245-248

59.

Nematollahi A, Decostere A, Pasmans F, Haesebrouck F (2003) Flavobacterium
psychrophilum infections in salmonid fish. J Fish Dis 26: 563-574

60.

Ostland VE, Lumsden JS, Macphee DD, Ferguson HW (1994) Characteristics of
Flavobacterium branchiophilum, the Cause of Salmonid Bacterial Gill Disease in Ontario.
J Aquat Anim Health 6: 13-26

61.

Hänninen M-l, Oivanen P, Hirvelä-koski V (1997) Aeromonas species in fish, fish-eggs,
shrimp and freshwater. Int J Food Microbiol 34: 17-26

62.

Carling PA (1992) The nature of the fluid boundary layer and the selection of parameters
for benthic ecology. Freshwater Biol 28: 273-284

63.

Yannarell AC, Triplett EW (2005) Geographic and Environmental Sources of Variation
in Lake Bacterial Community Composition. Appl Environ Microbiol 71: 227-239

64.

Sekiguchi H, Watanabe M, Nakahara T, Xu B, Uchiyama H (2002) Succession of
bacterial community structure along the changjiang river determined by Denaturing
Gradient Gel Electrophoresis and clone library analysis. Applied and Environmental
Microbiology 68: 5142-5150

53

CHAPTER 3: THE EFFECT OF TEMPERATURE AND WATER TYPE ON THE EGG
SURFACE MICROBIAL COMMUNITY AND HOST LIFE HISTORY TRAITS OF THE
LAKE STURGEON

Abstract
We investigated the process of microbial colonization on the egg surfaces of the Lake Sturgeon
(Acipenser fulvescens), a threatened fish species inhabiting the Great Lakes. Our previous studies
revealed that microbial community assemblages on the egg surfaces changed over the
developmental time of the egg (Chapter 2). To elucidate the factors influencing successional
changes, we designed a two factor experiment to analyze the effect of temperature (warm and
cold) and the aquatic microbes (filter/UV-treated and non-treated stream water) on the egg
surface microbial community assemblages throughout the developmental period. Genomic DNA
was extracted from the egg surface and the egg microbial communities were examined using 16S
rRNA gene based TRFLP and clone libraries. Microbial quantity on the egg surface was
determined by quantitative PCR (qPCR). Principle components analysis of T-RFLP data
revealed that the composition of the aquatic microbial community profoundly influenced the egg
associated microbial community assemblage during succession while temperature had a lesser
influence. The qPCR analysis showed that a significantly higher number of microbes were
attached to egg surfaces reared in non-treated water than in UV/filter-treated water (F1,210=35.67,
p<0.001). We also found a significantly higher egg mortality for eggs reared in stream water
(77.4 ± 17.6%) compared to those reared in UV/filter-treated water (50.3 ± 14.7%) (F1,17=13.42,
p=0.002). Eggs reared in stream water also had smaller yolk sac to body area ratio of larvae at
54

hatch compared to those reared in UV/filter-treated water (F1,156 = 52.39, p <0.001). The
phenomenon may have occurred because embryos used yolk sac resources for defenses against
microbes that colonized the egg surfaces. Clone libraries revealed that certain microbes were
associated with eggs that had a higher egg survival rate, including Acidovorax spp. and Massilia
spp., making them good candidates for probiotics. The results suggest that egg surface microbial
communities vary as a function of environmental factors and influence life history traits of the
host. This study also has significant implications for managing threatened host populations such
as the Lake Sturgeon inhabiting human-altered rivers, since it demonstrates the potential effect of
dams (which alter aquatic microbial community and temperature) on downstream host-microbe
interactions.

55

Introduction
Fish deposit eggs into stream water populated by aquatic microbes. Deposited eggs are
rapidly colonized by aquatic microbes, and a microbial community develops on egg surfaces
(Chapter 2). The interactions between microbes and host organisms play a significant role in
both a host’s life history and microbial life cycles. Fish eggs are an important focus of study
because the majority of mortality occurs during the egg stage of their life history [1]. Although
egg-associated microbes can be pathogenic or mutualistic and even symbiotic, treating eggs
against microbes with antibiotic or formalin helps increase the survivability of fish eggs [2-3].
The interaction between microbes and eggs not only affects the mortality of eggs, but also
potentially affects other life history traits such as the size of larvae at hatch [4].
Microbe-host interactions are long-term in nature and have in many cases coevolved to
the point of reaching a fine balance, and perturbing the environmental covariates will disrupt
their mutual relationships. The effect of such disturbances on microbe-host interactions is a
particularly important topic to examine with respect to aquatic systems. For example, a previous
study found that changes in temperature affected microbial communities associated with frog
hosts, which subsequently affected host mortality [5]. Damming streams has been shown to
change aquatic microbial community structure [6], but the effect of such disturbance on the
downstream host - microbe interactions is unknown.
Investigation into the effects of environmental perturbations such as damming on hostmicrobe interactions is particularly important for the threatened fish species Lake Sturgeon
(Acipenser fulvescens). Lake Sturgeon populations have decreased drastically over the past 100
years due to anthropogenic activities such as overfishing and dam construction [7]. One such
population is the Black Lake population in Michigan [8-9]. Spawning habitats have degraded in

56

quality since the construction of a dam in 1949 on the Upper Black River, which is the sole
spawning stream for the Black Lake population [9]. Lake sturgeon fish eggs serve as hosts for
diverse aquatic microbial communities (Chapter 2). While our previous research revealed that
microbial communities develop on the egg surfaces during the incubation period, many questions
remain regarding the interaction between microbial communities and their egg hosts.
The objective of this study was to determine the effect of variation in environmental
factors on the interaction between microbes and Lake Sturgeon egg hosts. We hypothesized that
(i) conditions of the rearing environment (aquatic microbial community composition and
temperature) affect the egg surface microbial communities and (ii) microbial communities on the
egg surfaces have an effect on egg mortality and other life history traits (larval size at hatch). To
test these hypotheses, we designed a two factor experiment to analyze the effect of water type
(filter/UV treated and non-treated stream water) and temperature (warm and cold) on the egg
associated microbial community assemblages throughout the incubation period, and their
subsequent impacts on host life history traits. This study has significant implications for
understanding microbe-host-environmental covariates interactions and for managing threatened
host populations such as the Lake Sturgeon inhabiting human-altered rivers, since it
demonstrates the potential effect of dams (which alter aquatic microbial communities and
temperature) on downstream ecosystems.

Methods
Study site
This experiment was conducted in a streamside hatchery located at the riverside of the
Upper Black River in Michigan, which is the sole spawning stream for the Black Lake
population of the Lake Sturgeon. The experiments were conducted in May 2009 during a part of
57

the annual spawning season. Stream water was pumped up from the spawning stream and large
particulate matter in the stream water was removed using sock filters before the filtered stream
water was gravity fed to the hatchery system. Both female and male gametes were collected from
spawning adult sturgeon in the Upper Black River. The collected gametes were fertilized and
reared in the hatchery under different treatment conditions.

Experimental design
The experiment was composed of two factors, water types and temperature of the egg
rearing environment. There were two levels for each factor- filter/UV-treated and non-treated for
o

o

water types, and cold (12 ± 1 C) and warm (18 ± 1 C) for temperature, respectively. The
experiment consisted of a total of four different treatments (warm/treated, warm/non-treated,
cold/treated, cold/non-treated). The manipulation of aquatic microbial communities and quantity
was conducted using a water treatment system which consisted of a 50µm filter cartridge
followed by a UV lamp (25 Watt, Emperor Aquatics, Inc). Under this treatment, microbes
associated with particulates larger than 50 µm and sensitive to UV treatment were selectively
removed from the stream water microbial community (Chapter 5).
For each treatment, approximately 200 gametes from a single female were fertilized with
1mL of milt from a single male on a sterile polyethylene mesh screen in a heath tray filled with
1L of either non-treated or filter/UV treated water . Gametes were left to sit for 30 to 40 minutes
until they were fertilized and adhered to the bottom polyethylene mesh screen, which served as
the bottom substrate for embryos throughout incubation. Eggs were then reared in their
respective treatments until they hatched. The study was replicated with five different
female/male combinations (family codes: BC, CE, DG, EI, and FK).
58

Egg samples were collected for DNA extraction and subsequent community analysis at
six different time points for each treatment. For each egg sampling event, 10 live eggs were
collected and placed in a sterile 2mL eppendorf tube that was filled with 80% ethanol and stored
o

at 4 C. For warm temperature treatments (warm/treated, warm/non-treated), egg samples were
collected at Day 0 (1 hour post fertilization) and Days 1, 2, 3, 4, and 5 post fertilization. All eggs
from warm treatments hatched at Day 6 in this experiment. For cold treatments (cold/treated,
cold/non-treated), egg samples were collected at Day 0 (1 hour post fertilization), and Days 2, 4,
6, 8, and 10 post fertilization. All eggs from the cold treatments hatched at Day 11, 12 or 13. The
o

development of the embryo was known to be accelerated at warm temperature (18 ± 1 C)
o

approximately by two-fold relative to the cold water treatment (12 ± 1 C) [10]. With this
sampling scheme, the developmental stages of the embryo were expected to be synchronized
between the two water treatments.

Mortality assessment
The numbers of dead eggs in each treatment group were recorded on a daily basis for
warm treatments and every other day for cold water treatments throughout the incubation period.
The arrest of embryonic development was determined by visual observation of developmental
stages of embryos relative to reference [11]. All dead eggs were removed from the incubation
tray at detection. The number of successful hatches for each treatment was recorded, and egg
mortality for each treatment was calculated as follows; egg mortality = total number of dead eggs
/ (total number of dead eggs + total number of hatches). The effect of water treatment and

59

temperature on egg mortality was assessed using a general linear model using the “lm” function
in the R software version 2.10.0 [12].
DNA extraction and TRFLP analysis
A total of 72 samples (3 families: CE, DG, EI, 4 treatments, 6 time points) were
processed for DNA extraction and subsequent community analysis using Terminal Restriction
Fragment Polymorphism (TRFLP). Microbial community genomic DNA was extracted from the
surfaces of 8 eggs per sample using the PowerSoil

TM

Kit (MO BIO Laboratories Inc., CA)

according to the manufacturer’s protocol. 16S rRNA gene based TRFLP was performed to
characterize microbial community structure [13-14]. The detailed PCR amplification procedures
for TRFLP were described in Chapter 2. The purified PCR products were digested with MspI
(Gibco BRL). Two technical replicates of each of the digested DNA samples were sent to
Michigan State University’s sequencing facility and the DNA fragments were separated on an
ABI 3100 Genetic Analyzer automated sequencer (Applied Biosystems Instruments, Foster City,
CA) in GeneScan mode. The sizes and abundance (peak height) of the terminal restriction
fragments (TRFs) were calculated using GeneScan 3.7. Each terminal fragment corresponds to a
phylotype, and peak height indicates relative abundance of a phylotype. In order to align TRFs
across egg samples from different treatments, the TRFLP profiles were processed with T-Align
software (http://inismor.ucd.ie/~talign/index.html) and the output of T-Align was used for the
microbial community analysis. Principle component analysis (PCA) was performed on TRFLP
data from the egg samples exposed to different treatments in order to elucidate underlying
patterns across samples. PCA was conducted using the “prcomp” function of the R software
version 2.10.0 [12].

60

16S rRNA gene Clone Library analysis
A total of four samples (CE family; warm/treated Day 1 and Day 4, and warm/nontreated Day 1 and Day 4) were subjected to a clone library to identify microbial species present
on the egg surface in two different water types at two different time points. 16S rRNA gene of
the extracted community DNA was amplified using 27F (5’ – AGA GTT TGA TCM TGG CTC
AG – 3’) and 1389R (5’-ACG GGC GGT GTG TAC AAG-3’). The PCR conditions were the
same as those used for TRFLP. PCR amplicons were purified and ligated into a vector plasmid
pCR2.1 with kanamicin resistance using a TOPO cloning kit (Invitrogen, Carlsbad, CA) and the
vector plasmid was transformed into competent E.coli cells. Vector bearing E.coli clones were
isolated on LB kanamicin plates with X-gal. A total of 96 white colonies per sample were picked
o

and inoculated in LB broth with kanamicin and grown overnight at 37 C. The insertion of the
amplicon into the vector plasmid was confirmed using M13 primers. Broth cultures were sent to
Michigan State University’s sequencing facility and the cloned 16S rRNA gene was sequenced
using a 27F primer. The sequences of the clone library were identified using RDP pipeline [15]
and the microbial community structures of different samples were compared at the genus level.

Quantitative PCR analysis
Microbial loads of the same 72 samples were determined by performing quantitative PCR
(qPCR) with SYBR green. The qPCR was performed using universal bacterial primers 331F (5’TCCTACGGGAGGCAGCAGT-3’) and 519R (5’- CGTATTACCGCGGCTGCT -3’) targeting
the conserved sequences within the 16S ribosomal RNA gene, as previously described [16]. The
qPCR reaction was conducted in a 25µL reaction volume, containing 3µL template DNA,
12.5uL mastermix (2x) (SABIoscience, MD: mixture of DNA polymerase, buffers, and SYBR
61

green), and 0.16µM of each primer in the final concentration. PCR was performed using the iQ5
(Bio-Rad) thermal cycler according to the following protocols: an initial denaturation at 94°C for
o

5 min followed by 43 cycles of denaturation at 94 °C for 30 s, annealing at 60 C for 30 s, and
extension at 72°C for 30 s. Fluorescence signals were detected at the end of the extension for
each cycle.
A standard curve for the relationship between 16S rRNA gene copy number and cycle
threshold (Ct) values were constructed using a series of dilutions of the bacterial genomic DNA
Flavobacterium johnsoniae ATCC 17061 that is known to have six 16S rRNA gene copies in its
genome. The standard curves were constructed over 20 times and the PCR efficiency of the
standard curves was found to range from 87% to 101%. For determination of microbial quantity
of the samples, a series of dilutions were made for each sample, and the triplicates of each
dilution were PCR amplified along with the standard. The Ct values for both the standard and
samples were determined at 600 relative fluorescence unit (RFU) where the relationship between
RFU and Ct was linear in the log-transformed view. The PCR efficiency of each sample was
comparable to that of the standard. The quantity of the 16S rRNA gene copy of each sample was
determined by substituting the Ct value of one of the sample dilutions into the equation of the
standard curve and multiplying it by the dilution factor.

Larval size analysis
Immediately after the hatching of eggs, the larvae were anesthetized using MS-222. The
anesthetized larvae were photographed with a ruler as a size standard. The total length, total
body area, and yolk sac area of the larvae were determined from the images using ImageJ
software. Yolk sac area (YSA) to body area (BA) ratio was chosen for measurement because
62

they allowed us to examine the effect of microbes on the allocation of yolk resources during
embryogenesis. A large YSA to BA ratio indicates that resources in the yolk sac were not used
for body area and/or immune responses during embryogenesis and are available for somatic
growth following hatch. The effect of water treatment and temperature and their interaction term
on the larvae size was assessed using both a general linear model with fixed variables and a
linear mixed effect model with family effect as a random variable. The statistical significance of
the treatment effect on larvae size was confirmed after accounting for family effect as a random
variable using the mixed effect model. The linear mixed effect model was performed using
“lme” function in R version 2.10.0 [12].

Results
A total of 146 phylotypes were detected using TRFLP across the 72 samples. On average,
28.0 ± 8.2 phylotypes were detected per sample. With regard to the principal component analysis,
principal component 1 (PC1) captured variability due to egg developmental stage, specifically
separating immediately post fertilized (1 hour after fertilization) egg samples from the rest of the
samples (Figure 3.1). The effect of water type on the egg associated microbial community was
captured by PC2, which separated the communities reared in filter/UV-treated water from those
reared in non-treated water. PC3 accounted for variation in the egg microbial communities due to
temperature (Figure 3.1).
PCA analysis performed on a subset of the data (only the warm treatment) showed that
the egg associated microbial communities collected from eggs reared in filter/UV treated water
was separated from that reared in non-treated water by PC1 (Figure 3.2). Variability in the egg
associated microbial community due to egg developmental stages was captured by PC2 (Figure

63

3.2). Samples collected immediately after fertilization (Day 0) were also separated from the rest.
The temporal trend was not random, but rather directional and corresponding to the egg
developmental stages (Figure 3.2).

(A)

Figure 3.1. Principal component analysis (PCA) plot of TRFLP data of 72 samples from
different treatments.
Panel (A) displays clustering due to temporal variation (PC1) and water treatment (PC2). “AF”
stands for after fertilization. “Early”, “Middle”, and “Late” indicate egg developmental stages.
“Early” corresponds to Day 1 for warm and Day 2 for cold, “Middle” corresponds to Day 2 and
Day 3 for warm and Day 4 and Day 6 for cold, and “Late” corresponds to Day 4 and Day 5 for
warm and Day 8 and Day 10 for cold. Panel (B) displays clustering due to water treatment (PC2)
and temperature (PC3).

64

(B)

Figure 3.1 (cont’d).

65

Figure 3.2. Principal component analysis (PCA) plot depicting the effect of water treatment
on egg microbial community at warm temperature.
PC1 and PC2 account for 16.6% and 11.5% of all TRFLP data variation, respectively.

66

The 16S rRNA gene based qPCR revealed that microbial quantity on the egg surfaces
increased in a log-linear fashion during the incubation period starting from 10
7.5

10

5.5

and ending at

16S rRNA gene copies per egg (Figure 3.3). Eggs reared in non-treated water had higher

microbial quantities than those reared in filter/UV treated water for all six time points tested
(Figure 3.3). The average quantity associated with eggs reared in non-treated water was
significantly higher than that reared in filter/UV treated water (F1,210=35.67, p<0.001, Figure
3.4). There was no significant effect of temperature on microbial quantity on the egg surfaces
(F1,210=0.178, p=0.674).
We identified a total of 47 genera using the 16S rRNA gene clone library (all except
singletons shown in Table 3.1). 8 genera were found only on or were more dominant on eggs
reared in non-treated water (Rhodobacter, Methylotenera, Polynucleobacter, Rhodoferax,
Leptothrix, Rheinheimera, Flectobacillus, and Flavobacterium). We also identified 7 genera that
were found only on or were more dominant on eggs reared in filter/UV-treated water
(Sphingobium, Massilia, Pseudorhodoferax, Pseudomonas, Acidovorax, Pelomonas, and
Aquabacterium).
Manipulation of aquatic microbial community using filter/UV treatment influenced egg
mortality. Mortality of eggs reared in filter/UV-treated water (50.3%) was significantly lower
than that reared in non-treated water (77.4%) (Figure 3.5, F1,17=13.42, p=0.002), whereas the
temperature of the rearing environment did not significantly affect egg mortality (F1,17 = 0.308,
p=0.59). There was also a family effect on egg mortality as evidenced by a moderately different

67

mortality of family CE from others (p = 0.15, 0.19, 0.09, and 0.20 for CE-DG, CE-EI, CE-FK,
and CE-BC, respectively). There was no significant difference in egg mortality

Figure 3.3. Positive linear relationship between time (days post-fertilization) and microbial
quantity present on egg surfaces.
Microbial quantity was measured using quantitative PCR (qPCR) on 16S rRNA gene copies.
Eggs reared in non-treated water had higher microbial quantity than that reared in treated water
at all time points.

68

Table 3.1. 16S rRNA gene-based clone library for the two different treatments at two
different time points for the CE family.
Class
Cyanobacteria
Alpha
Alpha
Alpha
Alpha
Alpha
Alpha
Alpha
Alpha
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Beta
Gamma
Gamma
Sphingobacteria
Flavobacteria

Genera
unclassified_Chloroplast
Caulobacter
Rhodobacter
Rhizobium
unclassified_Rhizobiales
Sphingomonas
Novosphingobium
Sphingobium
unclassified_Sphingomonadaceae
Methylophilus
Methylotenera
unclassified_Methylophilaceae
Polynucleobacter
Massilia
Hydrogenophaga
Acidovorax
Curvibacter
Rhodoferax
Pseudorhodoferax
Pelomonas
unclassified_Comamonadaceae
Aquabacterium
Leptothrix
Paucibacter
unclassified_Burkholderiales_incerta
unclassified_Burkholderiales
e_sedis
Rheinheimera
Pseudomonas
Flectobacillus
Flavobacterium

WN1
0.00
2.20
2.20
4.40
0.00
0.00
6.59
0.00
2.20
2.20
4.40
4.40
1.10
1.10
1.10
1.10
4.40
0.00
0.00
2.20
3.30
2.20
2.20
1.10
1.10
15.38
5.49
1.10
8.79
9.89

WF1
2.08
1.04
0.00
4.17
1.04
0.00
5.21
1.04
2.08
0.00
0.00
0.00
0.00
11.4
0.00
10.4
1.04
0.00
0.00
7.29
2.08
7.29
0.00
5.21
9.38
25.0
0.00
2.08
0.00
0.00

WN4
0.00
0.00
4.40
2.20
5.49
3.30
10.9
1.10
5.49
0.00
1.10
9.89
2.20
0.00
1.10
5.49
4.40
5.49
0.00
0.00
3.30
1.10
2.20
1.10
10.9
3.30
1.10
1.10
0.00
3.30

WF4
0.00
0.00
0.00
3.26
5.43
2.17
3.26
7.61
6.52
0.00
0.00
0.00
0.00
3.26
1.09
8.70
4.35
0.00
2.17
4.35
4.35
3.26
0.00
0.00
16.3
15.2
0.00
5.43
0.00
0.00

Numbers represent relative abundance in each community. Warm (W) rearing temperature,
Filter/UV (F) or Non-treated (N) at different developmental stages: Day 1 (1) or Day 4 (4).
Genera that showed marked differences in relative abundance between treated and non-treated
water are highlighted. Singletons are not included in the table.

69

Figure 3.4. The average microbial quantity associated with eggs as measured using qPCR.
Eggs reared in non-treated water had significantly higher microbial load than those reared in
filter/UV treated water (F1,210=35.67, p<0.001). Egg samples from both cold and warm water
treatments at all time points were included for this analysis.

70

Figure 3.5. The effect of water treatment on egg mortality.
Sturgeon eggs reared in non treated water had significantly higher egg mortality than those
reared in treated water (F1,17 = 13.42. p = 0.002). The treated group had 10 replicates and the
non-treated group had 9 replicates.

71

among families BC, DG, EI, and FK.
Eggs reared in cold water had significantly smaller yolk sac area to body area ratio than
those reared in warm water (F1,332 = 448.8, p <0.001, Figure 3.6). The effect of the aquatic
microbial communities on larvae size was dependent on temperature. Although we did not see a
significant difference in larvae size between the two water types for those eggs reared in cold
temperature, we did find a significant difference in larvae size between the two water types at
warm temperature (F1,156 = 52.39, p <0.001, Figure 3.6). The effect of rearing environment on
the larvae size remained significant after accounting for the family effect on the larvae size as a
random variable.

72

Figure 3.6. The effect of temperature and water treatment on resource allocation.
There was a difference in larvae size at hatch between eggs reared in cold water and warm water
(F1,332 = 448.8, p <0.001). There was a significant difference in larvae size between the two
water types at warm temperature (F1,156 = 52.39, p <0.001).

73

Discussion
This study demonstrated the significance of both aquatic microbial communities
(manipulated by water treatment) and temperature on the assembly of egg-associated microbial
communities and egg mortality. The study contributes to the literature on this topic by
highlighting the complexity of host-microbe interactions and the potential for rearing
environments to impact that relationship. To our knowledge this is the first study to examine
such effects in the threatened Lake sturgeon. This study also contributes to an understanding of
the impact of microbes on the life history of the vulnerable population of this species.
Our study demonstrated that different microbial communities developed on the egg
surface when reared in different aquatic microbial community. This result suggests that the type
of aquatic microbes that eggs were exposed to during fertilization and the subsequent incubation
period was important in determining microbial community composition on the egg surfaces. This
in turn suggests that dispersal of microbes from the water column onto the egg surfaces was an
important process in explaining microbial community assembly on the egg surfaces.
One of the most significant results from the study was the fact that the egg associated
microbial communities sampled immediately after fertilization (Day 0) were significantly
different from the rest. The Day 0 microbial communities could have originated from the aquatic
microbial community via dispersal during fertilization [17-18] and populated the host eggs
through adhesion. The subsequent changes in the egg associated microbial community observed
by Day 1 post fertilization could be attributed to local deterministic processes such as the effect
maternally provisioned innate immunity on the egg microbial community [19-21]. Alternatively,
the Day 0 microbial community on the egg surface could have been mainly attributed to
maternally provisioned community structure (if we consider unfertilized eggs to be non-sterile).

74

These two alternative hypotheses will be further explored using data on both water microbial
communities and unfertilized egg communities in forthcoming chapters (Chapter 4 and Chapter
5).
PCA analysis revealed a strong temporal variation within each water type (filter/UV or
non-filter). Microbial community on the egg surfaces changed directionally throughout the
embryogenesis, suggesting divergence from the initial community throughout embryogenesis.
The observed temporal variations could be derived from several factors including changes in
metabolites on the egg surfaces [22-23], changes in lysozyme from maternally provisioned to
egg secreted [24-25], or microbe-microbe interaction which we investigated in another chapter
(Chapter 6). We do not believe that the changes during embryogenesis can be attributed to
changes in the surrounding water microbial communities, since our rearing conditions (recirculating water) should have kept the water microbial communities relatively constant.
In the principal component analysis using the entire TRFLP data set, the first three
components accounted for about 30% of the entire variation within the dataset. This suggests that
factors other than temporal variation, water type, and temperature which we tested here
influenced the microbial community structures on the egg surfaces.
In addition to the temporal trend observed in microbial community composition over the
course of embryogenesis, the increase we observed in microbial quantity over the same period
5

was also noteworthy. The fact that over 10 16S rRNA gene copies per egg were detected
immediately post fertilization (within 1 hour) suggests that microbial colonization on the egg
surfaces during fertilization is rapid. There was an effect of water treatment on the quantity of
the egg associated microbes. Eggs fertilized and reared in non-treated water had significantly
higher microbial quantity on eggs than those fertilized and reared in filter/UV treated throughout
75

the experiments. This difference in quantity could be attributed to the difference in water
microbial quantity (Chapter 5). This significant difference suggests that dispersal of microbes
from the water column onto the egg surface is an important process in explaining egg surface
microbial population size.
We found that not only quantity but also quality of microbial community on the egg
surfaces were different between the samples reared in the two different water types. The
significance of this result can best be appreciated by considering the lower egg mortality that we
observed in filter/UV treated water compared to non-treated water. Microbes that were dominant
in non-treated water including genus Flavobacterium may have had negative effect on egg
survival. Flavobacterium is a genus known to include fish pathogens [26-28]. Microbes that
were only prevalent in filter/UV-treated water including genus Acidovorax could have had a
beneficial effect on egg survival. Acidovorax is a genus known to include a symbiont to
earthworms [29]. The significance of these putative pathogens and symbionts will be further
explored in Chapters 4. However, it is also important to note that the difference in egg mortality
between the two treatments could also be attributed to a difference in microbial quantity on egg
surfaces between the two treatments. The large microbial load on the egg surface alone could
prevent diffusion of oxygen to eggs and suffocate them [30].
Another significant result of this study was that temperature also affected the microbial
communities on the egg surfaces. However, our PCA analysis revealed that the effect of
temperature on the egg community structure was smaller relative to the effect of water treatment
on the egg microbial community. We found that temperature did not affect egg mortality,
although water temperature significantly changed the duration of the incubation period, and also
the size of larvae at hatch. The significantly smaller yolk sac area to body area ratio observed in

76

cold water-reared eggs compared to warm water-reared eggs is striking, a phenomenon termed
the “temperature-size rule” [31-32]. This result suggests that more resources were allocated to
body area for eggs reared under cold temperature at the expense of yolk sac area. The decrease in
yolk resources is significant because it may affect post-hatch development and feeding timing of
the larvae, which in turn could affect the probability of post-hatch survival. This result also has
implications for damming of the Lake Sturgeon habitat, since dams can affect downstream water
temperature, dependent on whether the water is released from the surface (epilimnion) or bottom
(hypolimnion) [33], which our results suggest may affect the microbial community on the egg
surfaces, incubation periods, and size at hatch.
Under cold temperature, the effect of rearing water types on yolk sac area to body area
ratio was negligible, since the dominant effect of cold temperature masked the effect of water
types. The fact that yolk sac area to body area ratio was smaller in eggs reared in non-treated
water compared to treated water at warm temperature suggests that the microbial quantity or
composition on the egg surface affected the yolk resource use by embryos.
In conclusion, this study demonstrated that rearing environments can affect both egg
associated microbial communities and host life history traits. This study demonstrates for the
first time that the effect of water treatment on improving Lake Sturgeon egg survival can be
attributed to a decrease in microbial quantity and/or change in the microbial community
composition on the egg surfaces. This study also provides evidence that rearing temperature
affects the egg microbial community assembly and colder temperatures can result in a smaller
yolk sac area to body area ratio. We also identified putative symbionts (e.g. Acidovorax spp. and
Massilia) associated with low egg mortality, which will help guide further studies on
identification of microbes for probiotic treatment. Our study provides management implications

77

for conserving Lake Sturgeon populations by suggesting that damming streams can alter aquatic
microbial community and temperature, which in turn can alter the microbial communities on
sturgeon eggs and life history of the sturgeon.

78

References

79

References

1.

Forsythe PS (2010) Exogenous correlates of migration, spawning, egg deposition and egg
mortality in the lake sturgeon (Acipenser fulvescens). Ph.D. Dissertation. Department of
Fisheries and Wildlife. Michigan State University. #3417681. pp191

2.

Marques A, Ollevier F, Verstraete W, Sorgeloos P, Bossier P (2006) Gnotobiotically
grown aquatic animals: opportunities to investigate host-microbe interactions. J Appl
Microbiol 100: 903-918

3.

Barnes ME, Gabel AC, Cordes RJ (1999) Bacterial populations during inland fall
chinook salmon egg culture in vertical-flow tray incubators. N Am J Aquacult 61: 252257

4.

Uller T, Sagvik J, Olsson M (2009) Pre-hatching exposure to water mold reduces size at
metamorphosis in the moor frog. Oecologia 160: 9-14

5.

Harris RN, Brucker RM, Walke JB, Becker MH, Schwantes CR, Flaherty DC, Lam BA,
Woodhams DC, Briggs CJ, Vredenburg VT, Minbiole KPC (2009) Skin microbes on
frogs prevent morbidity and mortality caused by a lethal skin fungus. ISME J 3: 818-824

6.

Sekiguchi H, Watanabe M, Nakahara T, Xu B, Uchiyama H (2002) Succession of
bacterial community structure along the changjiang river determined by Denaturing
Gradient Gel Electrophoresis and clone library analysis. Applied and Environmental
Microbiology 68: 5142-5150

7.

Adams PB, Grimes C, Hightower JE, Lindley ST, Moser ML, Parsley MJ (2007)
Population status of North American green sturgeon, Acipenser medirostris. Environ Biol
Fish 79: 339-356

8.

Smith KM, Baker EA (2005) Characteristics of spawning lake sturgeon in the Upper
Black River, Michigan. N Am J Fish Manage 25: 301-307

9.

Baker EA, Borgeson DJ (1999) Lake Sturgeon abundance and harvest in Black Lake,
Michigan, 1975-1999. N Am J Fish Manage 19: 1080-1088

10.

Wang Y, Binkowski F, Doroshov S (1985) Effect of temperature on early development of
white and lake sturgeon, Acipenser transmontanus and A. fulvescens. Environmental
Biology of Fishes 14: 43-50

11.

Colombo RE, Garvey JE, Wills PS (2007) A guide to the embryonic development of the
shovelnose sturgeon (Scaphirhynchus platorynchus), reared at a constant temperature. J
Appl Ichthyol 23: 402-410

80

12.

R Development Core Team (2009) R: A launguage and environment for statistical
computing. R foundation for Statistical Computing, Vienna, Austria.

13.

Liu W, Marsh T, Cheng H, Forney L (1997) Characterization of microbial diversity by
determining terminal restriction fragment length polymorphisms of genes encoding 16S
rRNA. Applied and Environmental Microbiology 63: 4516-4522

14.

Marsh TL (1999) Terminal restriction fragment length polymorphism (T-RFLP): An
emerging method for characterizing diversity among homologous populations of
amplification products. Current Opinion in Microbiology 2: 323-327

15.

Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS,
McGarrell DM, Marsh T, Garrity GM, Tiedje JM (2009) The Ribosomal Database
Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res 37:
D141-D145

16.

Smith-Vaughan H, Byun R, Nadkarni M, Jacques N, Hunter N, Halpin S, Morris P,
Leach A (2006) Measuring nasal bacterial load and its association with otitis media.
BMC Ear, Nose and Throat Disorders 6: 10

17.

Finlay BJ, Clarke KJ (1999) Ubiquitous dispersal of microbial species. Nature 400: 828828

18.

Dumbrell AJ, Nelson M, Helgason T, Dytham C, Fitter AH (2009) Relative roles of niche
and neutral processes in structuring a soil microbial community. ISME J 4: 337-345

19.

Kudo S, Inoue M (1989) Bacterial action of fertilization envelope extract from eggs of
the fish Cyprinus carpio and Plecoglossus altivelis. Journal of Experimental Zoology 250:
219-228

20.

Kudo S, Teshima C (1991) Enzyme activities and antifungal action of fertilization
envelope extract from fish eggs. Journal of Experimental Zoology 259: 392-398

21.

Kudo S (2000) Enzymes responsible for the bactericidal effect in extracts of vitelline and
fertilisation envelopes of rainbow trout eggs. Zygote 8: 257-265

22.

Chadwick T, Wright P (1999) Nitrogen excretion and expression of urea cycle enzymes
in the atlantic cod (Gadus morhua l.): a comparison of early life stages with adults. The
Journal of Experimental Biology 202: 2653-2662

23.

Steele SL, Chadwick TD, Wright PA (2001) Ammonia detoxification and localization of
urea cycle enzyme activity in embryos of the rainbow trout (Oncorhynchus mykiss) in
relation to early tolerance to high environmental ammonia levels. J Exp Biol 204: 21452154

24.

Fraune S, Augustin R, Anton-Erxleben F, Wittlieb J, Gelhaus C, Klimovich VB,
Samoilovich MP, Bosch TCG (2010) In an early branching metazoan, bacterial

81

colonization of the embryo is controlled by maternal antimicrobial peptides. Proceedings
of the National Academy of Sciences of the United States of America 107: 18067-18072
25.

Fraune S, Augustin R, Bosch TCG (2011) Embryo protection in contemporary
immunology: Why bacteria matter. Communicative and Integrative Biology 4: 369-372

26.

Nematollahi A, Decostere A, Pasmans F, Haesebrouck F (2003) Flavobacterium
psychrophilum infections in salmonid fish. J Fish Dis 26: 563-574

27.

Agarwal S, Hunnicutt DW, McBride MJ (1997) Cloning and characterization of the
Flavobacterium johnsoniae (Cytophaga johnsonae) gliding motility gene, gldA. Proc Natl
Acad Sci USA 94: 12139-12144

28.

Stringer-Roth KM, Yunghans W, Caslake LF (2002) Differences in chondroitin AC lyase
activity of Flavobacterium columnare isolates. J Fish Dis 25: 687-691

29.

Schramm A, Davidson SK, Dodsworth JA, Drake HL, Stahl DA, Dubilier N (2003)
Acidovorax-like symbionts in the nephridia of earthworms. Environ Microbiol 5: 804809

30.

Komar C, Turnbull JF, Roque A, Fajer E, Duncan NJ (2004) Effect of water treatment
and aeration on the percentage hatch of demersal, adhesive eggs of the bullseye puffer
(Sphoeroides annulatus). Aquaculture 229: 147-158

31.

Angilletta MJ, Steury TD, Sears MW (2004) Temperature, Growth Rate, and Body Size
in Ectotherms: Fitting Pieces of a Life-History Puzzle. Integr Comp Biol 44: 498-509

32.

Walters RJ, Hassall M (2006) The temperature-size rule in ectotherms: may a general
explanation exist after all? (vol 167, pg 510, 2006). American Naturalist 167: 775-775

33.

King J, Cambray JA, Dean Impson N (1998) Linked effects of dam-released floods and
water temperature on spawning of the Clanwilliam yellowfish Barbus capensis.
Hydrobiologia 384: 245-265

82

CHAPTER 4: THE EFFECT OF INOCULATION OF PUTATIVE SYMBIONTS ON
LAKE STURGEON EGG SURFACE MICROBIAL COMMUNITY ASSEMBLY AND
EGG MORTALITY

Abstract
Microbes and their hosts undergo complex interactions and may sometimes form
symbiotic relationships, where both the host and microbes benefit from being associated with
one another. Few studies have investigated this topic in fish egg hosts, but research is strongly
needed in order to identify potential symbionts that could improve egg survivorship for
threatened fish species. In this study, the effect of supplementation of putative symbionts during
fertilization on the egg microbial community assembly, egg microbial quantity, and egg
mortality was examined in the threatened fish species Lake Sturgeon (Acipenser fulvescens). The
experiment consisted of two treatment groups (eggs fertilized in the presence of a putative
symbiont Acidovorax sp. F19 and a putative pathogen Flavobacterium sp. B17) and two controls
(eggs fertilized with 0.2 µm filtered water and eggs fertilized with stream water). Egg samples
for community analysis were collected at two time points (immediately after fertilization and day
2 post-fertilization). Microbial community structure was examined using TRFLP and 454
pyrosequencing and microbial quantity was estimated using qPCR. Fertilization in the presence
of Acidovorax sp. F19 reduced egg mortality by 18% relative to those fertilized in stream water
(t = 1.98, p = 0.06). There was no difference in egg mortality between Flavobacterium treatment
and stream control. Analysis of microbial community composition immediately after fertilization
across the treatments indicated that Acidovorax sp. F19 successfully colonized egg surfaces and
became dominant in the community, whereas Flavobacterium sp. B17 was not as effective at
83

colonizing eggs. Despite the differences in initial microbial community compositions across the
treatments and controls, the egg associated microbial communities of the treatment and control
groups converged after two days of incubation in stream water. Microbial quantity on the egg
5

7

surfaces increased from 10 to 10 16S rRNA gene copies per egg by day 2, but there was no
significant differences in the egg associated microbial quantity between treatments and controls.
Early supplementation with Acidovorax sp. F19 resulted in a significant increase in yolk sac to
body area ratio of eggs (t116 = 2.47, p = 0.015). Acidovorax sp. F19 was closely grouped with
one of the branches within beta proteobacteria clade found on unfertilized eggs, suggesting the
possibility of vertical transmission of this putative symbiont. Although we do not observe any
differences in microbial community structure and quantity at day 2, that significant effect of
Acidovorax supplementation on egg mortality and yolk sac to body area ratio provides strong
evidence that this microbe displays a symbiotic relationship with the Lake Sturgeon egg host.

84

Introduction
The word symbiosis originates from the Greek words “syn” and “biosis”, which mean
“living” and “together”. When the word is used to refer to the interaction between microbes and
hosts, it usually indicates a long-term mutual relationship between the two. There are no clear
criteria to determine a symbiotic relationship between hosts and microbes, unlike the well
established criteria for definition of pathogens [1]. However, symbiotic relationships observed
between microbes and hosts in nature cover a wide range of spectra, including provision of
essential nutrients to the host [2-3], protection of the host from pathogenic organisms [4], and
conferral of a certain evasion ability to the host [5], all of which suggest that the association of
hosts with symbionts increases fitness of the hosts.
However, the topic of symbiosis with microbes is rarely investigated in fish egg hosts,
despite the fact that microbial symbionts may potentially significantly affect egg development
and survival. Such research is needed for the Lake Sturgeon (Acipenser fulvescens), a threatened
fish species inhabiting the Great Lakes, which deposits eggs into stream water populated by
aquatic microbes. Deposited eggs rapidly become colonized by aquatic microbes and a microbial
community develops on the egg surface. These egg-associated microbes can be pathogenic,
commensal, or even symbiotic.
The existence of symbiotic microbes for fish eggs may seem unlikely, considering the
fact that the interaction between microbes and eggs could be seen as too short in duration to
establish long-term mutual relationship. However, it is possible that some microbial species
protect eggs during embryogenesis and subsequently colonize the larvae and establish a longterm relationship thereafter [6]. One way of establishing such relationship is via horizontal
transfer [7]. Because adults of the same sturgeon population spawn in the same river for

85

thousands of years, aquatic microbe–sturgeon interaction can be viewed as long-term and may
have coevolved. The sturgeon eggs may have adapted to interact with certain aquatic microbial
species in a mutualistic way by controlling their immune response in a way that symbionts can
tolerate, a phenomenon which was observed in other hosts [8-11].
Another way of establishing a symbiotic relationship with microbes is via vertical
transmission [6]. A long-term relationship between host and microbes can be achieved by
directly transmitting symbionts from parents to offspring. For example, Acidovorax sp., a
symbiont of the earthworm, is transmitted from the parent into the egg capsule during
fertilization. Acidovorax sp. then colonizes the embryonic duct, excludes other microbes from
this area, and migrates to the nephridium (an excretory organ) of the earthworm offspring [6].
Vertical transmission of microbial symbionts from parents (or a parent) to offspring is also
documented in the medical leech [12], nematodes [13], insects [14-15], and sponges [16-18].
Since the Lake Sturgeon is a long-lived animal [19], the possible existence of vertical
transmission of symbionts to eggs could be a great benefit for both symbionts and eggs.
Previously, we experimentally identified a putative pathogen (Flavobacterium sp.) and a
symbiont (Acidovorax sp.) for Lake Sturgeon eggs based on the correlation between their
dominance on the egg surface and egg mortality under different water treatments (Chapter 3).
We also isolated these putative pathogen and symbiont from the egg surfaces using R2A media
(Chapter 6). In this study, we sought to elucidate the effect(s) of a putative symbiont (Acidovorax
sp.) and pathogen (Flavobacterium sp.) on egg mortality by experimentally supplementing the
eggs with the isolate during fertilization. Pathogenesis and symbiosis of these two isolates were
assessed based on comparing the difference in egg mortality between a treatment and a control
(fertilization with stream water). We also confirmed the association of these strains with the egg

86

by monitoring changes in microbial community structure with T-RFLP and measured the level of
colonization with Q-PCR on the surfaces of eggs exposed to different treatments and controls.
We hypothesized that a probiotic would protect eggs by altering microbial community and/or
microbial quantity on the egg surfaces. Identifying symbionts to the Lake Sturgeon eggs is
urgently needed, since populations of this species are intensively threatened by anthropogenic
pressure including overfishing and dam construction [20-22].

Methods
Experimental design
Eggs were fertilized in four different conditions (two treatments and two controls): (1)
with a putative symbiont Acidovorax sp. F19 (2) with a putative pathogen Flavobacterium sp.
B17 (3) with 0.2 µm filtered water, and (4) with stream water. Fertilization in 0.2 µm filtered
stream water served as a negative control (no microbial impact during fertilization) and the
fertilization with stream water served as a control for reference level of mortality at hatchery.
For each of the two treatment conditions, one hundred eggs were fertilized in 0.2 µm filter
treated stream water using a 0.22 µm disposable 500 mL membrane filter (Corning Inc) but
6

containing the final concentration of 10 cfu/mL of either Acidovorax sp. F19 or Flavobacterium
sp. B17. We chose this range of concentration because it corresponds to average microbial
concentrations in freshwater [23]. To control the concentration of inocula, a growth curve (the
relationship between microbial cell density and absorbance) was constructed for each isolate by
measuring optical density of broth culture at 600 µm using a spectrometer and enumerating the
viable cell counts of the broth culture on R2A plate medium. The detailed fertilization process
was described in Chapter 3, the only difference being that eggs were fertilized in a smaller tray
87

(500 mL volume) with 0.1 mL milt per 500 mL liquid. After fertilization, Acidovorax sp. F19
and Flavobacterium sp. B17 coated eggs and the two controls were reared in stream water at 18
o

to 19 C, which was known to cause about 78% egg mortality from previous years of experiments
(Chapter 3). We had a total of seven replicates (gametes from seven different pairs of parents)
for this experiment.
Egg samples were collected at two time points (immediately after fertilizations, and day2
post fertilization) from the treatments and controls. The 2 time points were incorporated in order
to address our hypothesis that inoculation of putative pathogen or symbiont alters subsequent
microbial community structure on egg surfaces. Ten live eggs were collected for after
fertilization egg samples, and 5 live eggs were collected for Day 2 post fertilization egg samples.
o

Egg samples were rinsed in filtered stream water and stored in 80% ethanol at 4 C. Unfertilized
o

eggs from each family were also collected and stored in 80% ethanol at 4 C. Stream water
samples were collected at two time points (immediately before fertilizations, and Day 2 post
fertilization). For each sampling, 100 mL stream water was filtered with 0.22um filter membrane
o

and the filter membrane with microbes was stored in 80% ethanol at 4 C.

Assessment of egg mortality
The death of an egg was defined as the arrest of embryonic development, which was
determined by visual observation of developmental stages of embryos [24]. The number of dead
eggs was recorded for each treatment and control on a daily basis, and all dead eggs were
removed from the incubation tray upon detection. The number of successful hatches for each
treatment and control were also recorded. The cumulative egg mortality for each treatment and
88

control was calculated as follows: Egg mortality = total number of dead eggs / (total number of
dead eggs + total number of hatches). The cumulative egg mortality between treatments and
controls were then compared. The effect of treatment on egg mortality was assessed using both a
general linear model using the “lm” function and a contrast analysis between the treatment and
control using ”glht” in the R software version 2.10.0 [25].

DNA extraction and TRFLP analysis
A total of 27 egg microbial community samples (3 families, 2 treatments and 2 controls,
2 time points, plus 3 unfertilized eggs) and 2 water microbial community samples (stream water
collected at 2 time points) were processed for DNA extraction and subsequent community
analysis using Terminal Restriction Fragment Polymorphism (TRFLP). Three aseptically
harvested unfertilized eggs and stream water samples were included for comparison with eggassociated communities.
For all samples, microbial community genomic DNA was extracted using the
PowerSoil

TM

Kit (MO BIO Laboratories Inc., CA) according to the manufacturer’s protocol. For

egg samples, the DNA was extracted from the surfaces of 8 eggs per sample for immediately
after fertilization samples and extracted from the surfaces of 4 eggs per sample for Day 2 post
fertilization samples. Previous experiments have indicated a substantially denser community on
eggs after two days of incubation in stream water. For water samples, genomic DNA was
extracted from the filtered material.
16S rRNA gene based TRFLP was performed to characterize microbial community
structure [26-27]. The detailed PCR amplification procedures for TRFLP were described in
Chapter 2. The purified PCR products were subjected to enzyme digestion with HhaI (Gibco
89

BRL). Two technical replicates of each of the digested DNA samples were sent to Michigan
State University’s sequencing facility and the DNA fragments were separated on an ABI 3100
Genetic Analyzer automated sequencer (Applied Biosystems Instruments, Foster City, CA) in
GeneScan mode. The sizes and abundance (peak height) of the terminal restriction fragments
(TRFs) were calculated using GeneScan 3.7. Each terminal fragment corresponds to a phylotype,
and peak height indicates relative abundance of a phylotype. In order to align TRFs across egg
samples from different treatments, the TRFLP profiles were processed with T-Align software
(http://inismor.ucd.ie/~talign/index.html) and the output of T-Align was used for microbial
community analysis. In order to confirm the TRFs of Acidovorax sp. F19 and Flavobacterium sp.
B17, pure cultures of each isolate were subjected to TRFLP analysis. Principal component
analysis (PCA) was performed on the TRFLP data in order to elucidate underlying patterns
across samples. PCA was conducted using the “prcomp” function of the R software version
2.10.0 [25].

qPCR
Quantification of microbial communities of the same 27 egg microbial community
samples and 2 water samples was determined with quantitative PCR (qPCR) using SYBR green.
The qPCR was performed using the same protocol described in Chapter 3. A standard curve for
the relationship between 16S rRNA gene copy number and cycle threshold (Ct) values was
constructed using a series of dilutions of the bacterial genomic DNA Flavobacterium johnsoniae
ATCC 17061 that is known to have six 16S rRNA gene copies in its genome. The quantity of the
16S rRNA gene copy in each sample was determined by substituting the Ct value of one of the
sample dilutions into the equation of the standard curve and multiplying it by the dilution factor.

90

454 pyrosequencing analysis
A subset of the aforementioned samples including (1) stream water at pre-fertilization, (2)
stream water at day 2 post-fertilization, and (3) an egg sample fertilized and reared in stream
water at day 2 post-fertilization was submitted to pyrosequencing analysis. This subset was
chosen to determine which microbial genera/species occurred naturally in streams and whether
their concentrations differed on egg surfaces. The V3-V5 region of the 16S rRNA gene of the
extracted DNA (see section above) was sequenced using 454 GS FLX titanium platform (454
Life Science, Branford, CT) at the research facility at Baylor, Texas. Raw sequence reads were
processed using Ribosomal Database Project (RDP) pipeline [28] to sort the data by tag sequence,
to trim tag and primer sequences, and to filter out low quality sequences with a minimum quality
score of 20 (probability threshold of 0.01) and a minimum read length of 300bp. The taxonomy
of the filtered reads was assigned using RDP Classifier at a bootstrap threshold of 80% [29].
Microbial communities of the samples were then compared at the genus level.

Vertical transmission analysis using 16S rRNA gene cloning
We investigated possible vertical transmission of symbionts from parent to egg by
harvesting unfertilized gametes from a female in the year 2009 during the spawning season. The
microbial community of the aseptically harvested gametes was subsequently examined using 16S
rRNA clone library. Microbial community genomic DNA was extracted from the surfaces of 10
gametes using the PowerSoil

TM

Kit (MO BIO Laboratories Inc., CA) according to the

manufacturer’s protocol. 16S rRNA gene of the extracted community DNA was amplified using
27F (5’ – AGA GTT TGA TCM TGG CTC AG – 3’) and 1389R (5’-ACG GGC GGT GTG
91

TAC AAG-3’). The PCR conditions were the same as those used for TRFLP. PCR amplicons
were purified and cloned into E.coli cells via a vector plasmid pCR2.1 using a TOPO cloning kit
(Invitrogen, Carlsbad, CA). The detailed procedure was described in Chapter 3. A total of 48
clones were sequenced at the Michigan State University’s sequencing facility using a 27F primer.
The sequences of the clone library were identified using RDP pipeline [28]. Sequences of the 48
clones and the putative symbiont Acidovorax sp. F19 were aligned using the Ribosomal Database
Project (RDP) [28]. The phylogenetic relationships of the aligned sequences were inferred using
MEGA version 4.0 [30] by constructing a Neighbor-Joining tree [31] with the Maximum
Composite Likelihood method [32].

Larval size analysis
The effect of putative symbiont (Acidovorax sp. F19) on the ratio of yolk sac area (YSA)
to body area (BA) ratio was measured as well. This measurement permits the evaluation of
treatment on the allocation of yolk resources during embryogenesis. A larger YSA to BA ratio
indicates that resources in the yolk sac were less used for body area and/or immune responses
during embryogenesis and were available for somatic growth following hatch. We hypothesized
that if eggs are inoculated with a symbiont, an embryo would use less yolk sac resources and
thus have a larger yolk sac area to body area ratio. To test this, we compared the yolk sac area to
body area ratio of Acidovorax sp. F19 fertilized larvae to that of stream water fertilized.
Immediately after the hatching of eggs, the larvae were anesthetized using MS-222.
Twenty individuals per treatment (or control) per family were photographed with a ruler as a size
standard. Photos of new hatched larvae from each of 7 families were included for the analysis.
For each larvae, the total length, total body area, and yolk sac area were determined from digital

92

images using ImageJ software. The effect of Acidovorax sp. F19 inoculation was assessed using
a linear mixed effect model with family effect as a random variable [Model<-lme (YSA.to.BA ~
Treatment, random=~1|Family)]. The statistical significance of the treatment effect on larvae size
was determined after accounting for family effect as a random variable using the mixed effect
model. The linear mixed effect model was performed using the “lme” function in R version
2.10.0 [25].

Results
Mortality analysis
Eggs fertilized in the presence of Acidovorax sp. F19 had an average mortality of 26.3%,
which was about 18 percent lower than that fertilized in the stream water (44.6%), and the
difference was marginally significant (t24=1.98, p=0.06; box plot shown in Figure 4.1). There
was no significant difference in mortality between Acidovorax sp. F19 (26.3%) and eggs
fertilized with 0.2 µm filtered water (35.5%) (t24=1.00, p=0.33). Treatment with Flavobacterium
sp. B17 during fertilization did not significantly affect the egg mortality relative to the stream
water fertilized control (t24=0.27, p=0.79). Eggs fertilized with Flavobacterium sp. B17 had a
higher egg mortality than those in 0.2 µm filtered water, although the difference was not
statistically significant (t24=0.71, p = 0.48).

93

Figure 4.1. A box plot showing the effect of the treatments on egg mortality.
“Stream” corresponds to eggs fertilized in stream water, “Flavo” corresponds to eggs fertilized
with Flavobacterium sp. B17, “0.2 µm” corresponds to eggs fertilized in 0.2 µm filtered water,
and “Acido” corresponds to eggs fertilized with Acidovorax sp. F19.

94

Microbial community analysis using TRFLP
To confirm that treatments during the fertilization changed the egg associated microbial
community, we analyzed the egg surface microbial community using TRFLP. TRFLP revealed
that there was a difference in microbial community composition immediately after fertilization
across the treatments and controls. Acidovorax sp. F19 successfully adhered to the egg surfaces
and became dominant in the community (accounting for 60 to 65% of the entire community).
Despite being inoculated with the same concentration as Acidovorax sp. F19,
Flavobacterium sp. B17 did not colonize the egg surfaces effectively, accounting for only 10 to
15% of the entire community (Figure 4.2). Microbial communities on the surfaces of eggs
fertilized in Flavobacterium sp. B17 were similar to that of unfertilized eggs (Figure 4.2). Eggs
fertilized in 0.2 µm filtered water had the almost identical community structure to that of
unfertilized eggs (Figure 4.2). Microbial communities on eggs fertilized in stream water diverged
from the unfertilized egg microbial communities and approached the stream water microbial
communities, although they were not identical (Figure 4.2). Although the initial microbial
community compositions differed across the treatments and controls, the egg associated
microbial communities of the treatments and controls converged by Day 2 post-fertilization
(Figure 4.3). The converged microbial community assembly on the egg surface was a subset of
the stream water community, but the community structure was significantly different from
stream water microbial community (Figure 4.3).
These trends were also detected in a PCA plot constructed using the TRFLP data (Figure
4.4). Microbial communities of unfertilized eggs were clustered with those fertilized in 0.2 µm
filtered water and those fertilized in Flavobacterium sp. B17 (Figure 4.4). Microbial

95

communities fertilized in stream water diverged significantly from those of unfertilized eggs and
approached the microbial community structure of stream water (Figure 4.4). Day 2 egg samples

Figure 4.2. Community assembly of unfertilized eggs (Unf) and eggs immediately after
being fertilized (AF).
Fa-Fc corresponds to family identities. “Aci”, “0.2”, “Fla” and “Str” stand for Acidovorax
inoculated, 0.2 µm filtered water (negative control), Flavobacterium inoculated, and stream
water (control for reference mortality), respectively. Water BF corresponds to stream water
microbial community collected immediately before fertilization

96

Figure 4.3. Convergence of the egg microbial community structure at Day 2 postfertilization.
Data were obtained from family “a”. Labels on the x axis correspond to those used in Figure 4.2.
The letter codes “BF”, “AF”, and “D2” correspond to before fertilization, after fertilization, and
Day 2 post-fertilization, respectively. The egg microbial communities from different treatments
and controls converged at Day 2 post-fertilization, but diverged from the stream water microbial
community.

97

Figure 4.4. Principal component analysis (PCA) analysis using TRFLP data from both egg
and water samples.
Microbial communities of unfertilized eggs were clustered with those fertilized in 0.2 µm filtered
water and those fertilized with Flavobacterium sp. B17 at immediately after fertilization.
Acidovorax sp. F19 samples departed from unfertilized egg samples. Stream water fertilized egg
communities further diverged from unfertilized communities toward the stream water microbial
community structure. Day 2 communities were all similar to each other regardless of initial
treatment and were distinct from unfertilized communities and stream water communities.
Labeling used is the same as that in Figures 4.2 and 4.3.

98

were all clustered together and diverged from both unfertilized eggs and stream water microbial
communities (Figure 4.4). There was variation among the egg microbial communities fertilized
in the stream water. One of the stream water fertilized egg samples at immediately after
fertilization was similar to the Day 2 microbial community cluster, but the other two fell
somewhere in between unfertilized eggs and stream water.

Microbial quantity analysis using qPCR
The microbial quantity on the egg surfaces immediately after fertilization was about 10

5

16S rRNA gene copies per egg for Acidovorax sp. F19, Flavobacterium sp. B17, and stream
water fertilized eggs (Figure 4.5). The microbial quantity on eggs fertilized in 0.2 µm filtered
4

water had a concentration below the detection limit (less than 10 16S rRNA gene copies per
egg). There was no measurable difference in the quantity of microbes on the egg surfaces at the
log scale between the Acidovorax fertilized treatment and the stream water fertilized control.
5

7

Microbial quantity increased from 10 to 10 16S rRNA gene copies per egg from fertilization to
Day 2 post-fertilization for all treatments (Figure 4.5).

99

Figure 4.5. Microbial quantity on the egg surfaces of various treatments and controls at
different time points estimated using qPCR.
Notation used on the x axis is the same as that used in Figures 4.2 and 4.3.

Microbial community analysis using 454 pyrosequencing
To determine what microbial genus were selected for or against by eggs, we compared
stream water microbial community to the eggs surfaces microbial communities fertilized in the
same stream water using pyrosequencing data. Approximately 7000 reads per sample were
obtained from each sample and average sequence length for the pre-fertilization stream water

100

sample, Day 2 post-fertilization stream water sample, and Day 2 post-fertilization stream water
fertilized egg sample were 509bp, 501bp, and 504bp, respectively.
At the phylum level of analysis, both stream water samples were dominated by phyla
Bacteroidetes, Proteobacteria, Actinobacteria, Cyanobacteria, and Firmicutes (Table 4.1).

Table 4.1. Comparison of stream water microbial community and stream water fertilized
egg microbial community at the phylum level using 454 pyrosequencing data
Phylum
Proteobacteria
Bacteroidetes
Actinobacteria
Cyanobacteria
Verrucomicrobia
Firmicutes
OD1
Acidobacteria
Chloroflexi
Planctomycetes
Chlamydiae
Nitrospira
Gemmatimonadet
es
Chlorobi
TM7
Spirochaetes
WS3
Elusimicrobia
Fusobacteria
Armatimonadetes
Deinococcus
Unclassified
Total reads
Number of phyla

Water BF
3569
1736
615
184
108
102
91
62
23
17
10
9
8
6
5
4
3
2
1
1
1
452
7009
21

Phylum
Water D2 Phylum
Egg D2
Bacteroidetes
3880 Proteobacteria
6492
Proteobacteria
2778 Bacteroidetes
478
Actinobacteria
586 Actinobacteria
78
Verrucomicrobia
28 Cyanobacteria
30
Firmicutes
24 OD1
15
Cyanobacteria
17 Verrucomicrobiai
15
OD1
16 Firmicutes
5
Armatimonadetes
7 Deinococcus
3
Deinococcus
6 Acidobacteria
2
Chlamydiae
5 WS3
1
TM7
3 Planctomycetes
1
Acidobacteria
1
Chloroflexi
1

76
7428
13

*** “BF” denotes before fertilization and “D2” denotes Day 2 post-fertilization.

101

109
7229
11

However, the Day 2 egg surface microbial communities were noticeably different from the water
community. The numbers of reads of Bacteroidetes, Actinobacteria, Firmicutes were reduced by
80%, 88% 91%, respectively, on the egg surfaces compared to the average number in the stream
water (Table 4.1). The number of phylum Proteobacteria on the egg surfaces was double that of
the stream water (Table 4.1).
At the genus level, the Day 2 egg surface microbial community had a significantly lower
number of Flavobacterium compared to those in the water samples (approximately 80% lower,
Table 4.2). The numbers of Limnohabitan, Fluviicola, Polynucleobacter, Arcicella were also
lower on the egg surfaces compared to the water samples (Table 4.2). In contrast, the egg surface
microbial community had a significantly greater number of Acidovorax compared to that in the
water samples (approximately 4 times more, Table 4.2). The numbers of Albidiferax,
Rheinheimera, Novosphingobium, Roseateles, Caulobacter, and Hydrogenophaga were also
higher on the egg surfaces than the water samples (Table 4.2).

Vertical transmission of Acidovorax sp. F19
The clone library showed that the microbial community on eggs harvested aseptically
consisted of microbes from four distinct major clades, which corresponded to Bacteroidetes,
alpha, beta, and gamma proteobacteria. Acidovorax sp. F19 was closely grouped with one of the
branches within the beta proteobacteria clade found on the aseptically harvested unfertilized eggs
(Figure 4.6).

102

Table 4.2. Comparison of water and egg microbial communities at the genus level using 454
pyrosequencing data.
Genus
Limnohabitans
Flavobacterium
Polynucleobacter
Pseudomonas
Methylophilus
Bacillariophyta
Fluviicola
OD1_incertae_sedis
Albidiferax
Arcicella
Sediminibacterium
Opitutus
Massilia
Lishizhenia
Sphingomonas
Rhodobacter
Polaromonas
Ilumatobacter
Acidovorax
Clostridium sensu
Gp6
Acinetobacter
Cryptomonadaceae
Ohtaekwangia
Legionella
Rhizobacter
Rheinheimera
Aquabacterium
Brevundimonas
Sphaerotilus
Other classified
Unclassified genus
Total reads
Number of genera

Water
BF
1162
764
259
126
123
106
103
91
81
66
54
50
48
47
46
45
37
33
32
29
27
26
24
23
23
19
18
17
16
15
628
2871
7009
191

Genus
Flavobacterium
Limnohabitans
Arcicella
Fluviicola
Lishizhenia
Sediminibacterium
Pseudomonas
Polynucleobacter
Rheinheimera
Sphingobium
Acidovorax
Brevundimonas
Rhodobacter
Albidiferax
Rhizobium
Massilia
Sphingomonas
Caulobacter
Novosphingobium
OD1_incertae_sedis
Pedobacter
Hydrogenophaga
Achromobacter
Undibacterium
Aurantimonas
Prosthecobacter
Arthrobacter
Algoriphagus
Methylophilus
Bacillariophyta

Water
D2
2376
1221
275
211
176
162
128
112
88
80
79
74
62
54
40
34
29
25
21
16
16
14
12
12
12
12
11
11
11
11
221
1822
7428
108

Genus
Albidiferax
Rheinheimera
Flavobacterium
Acidovorax
Novosphingobium
Undibacterium
Roseateles
Caulobacter
Pseudomonas
Brevundimonas
Hydrogenophaga
Massilia
Sphingobium
Limnohabitans
Rhodobacter
Sphaerotilus
Pelomonas
Duganella
Flectobacillus
Aquabacterium
Naxibacter
Vogesella
Rhizobium
Ideonella
Rhizobacter
Methylophilus
Sphingomonas
Fluviicola
OD1_incertae_sedis
Leptothrix

Egg
D2
819
708
267
231
212
195
167
165
155
145
123
116
115
103
76
72
69
68
53
45
44
43
42
38
23
22
19
18
15
14
277
2770
7229
102

*** Only the top 30 genera in each community are shown. “BF” denotes before fertilization and
“D2” denotes Day 2 post-fertilization.
103

Figure 4.6. Potential vertical transmission of Acidovorax sp. F19.
Acidovorax sp. F19 was closely grouped with beta proteobacteria clones identified on aseptically
harvested unfertilized eggs. MA00196 represents the plate number of 48 clone library.

104

Larvae size analysis
We found that larvae treated with Acidovorax sp. F19 during the fertilization had a
significantly larger yolk sac area to body area ratio than those fertilized in stream water after
taking into account the family effect as a random variable in the mixed effect model (t116 = 2.47,
p = 0.015, Figure 4.7).

Figure 4.7. Box plot depicting the effect of Acidovorax sp. F19 treatment on yolk sac area to
body area ratio of eggs.
105

Discussion
This study adopted an integrative approach to further evaluate microbe-host interactions,
specifically those involving a putative symbiont and its Lake Sturgeon egg host. We gathered
solid evidence to suggest the existence of a symbiotic relationship between Acidovorax sp. F19
and the eggs, in addition to identifying some probable mechanisms to explain the occurrence of
this symbiosis. The significant effect of inoculation of Acidovorax sp. F19 to eggs during
fertilization on reducing the egg mortality and yolk sac resource use suggest that this microbe
may be promising for probiotic treatment in this threatened fish species.
One of the surprising results from this study was that the egg associated microbial
community of all treatments and controls converged by Day 2 post-fertilization. This result was
counter to our hypothesis that we would observe measurable differences in egg surface microbial
communities on Day 2 across treatments. This is probably due to the fact that the microbial
quantity increased by 100 fold from fertilization to Day 2 post-fertilization, such that a massive
amount of dispersal from the stream water onto the egg surfaces effectively masked the effect of
any treatment given at fertilization. It was also surprising that we did not detect any difference in
microbial quantity on the egg surfaces between eggs fertilized with Acidovorax sp. F19 and those
fertilized with stream water at Day 2 post-fertilization. These lines of evidence suggest that the
mechanism to explain the observed symbiotic relationship between Acidovorax sp. F19 and eggs
occurred by some pathway other than alteration of the microbial community structure and
microbial quantity.
One possible mechanism would be that Acidovorax did not induce a host immune
response. Eggs may possess a mechanism to recognize the surface molecules of microbes and
transduce signals that determine the host responses to the microbes, a phenomenon which was

106

found in other host tissues [8-11]. It is known that some symbionts do not induce host immune
response at attachment [10]. This potential hypothesis is consistent with our finding that the egg
yolk resources were not as heavily used upon colonization by Acidovorax sp. F19 relative to the
stream water control. However, this hypothesis does not explain the lower egg mortality we
observed with Acidovorax sp. F19 treatment, unless expression of immune response to microbes
alone negatively affected the host and increased host mortality.
Another possible explanation could be that the relationship between Acidovorax and the
host eggs was neutral. In fact, across various host-microbe interactions observed in nature, the
majority of microbes are neutral to their hosts [33-34]. This hypothesis is supported by the fact
that we did not observe statistically significant differences in egg mortality between eggs
fertilized in Acidovorax sp. F19 and those fertilized in 0.2 µm filtered water. This line of
evidence suggests that Acidovorax acted essentially as a no microbe effect. There are numerous
micropyles (structures through which a sperm enters into an egg) on each egg of various
sturgeon species [35]. It is possible that some microbes could have entered into the inert area of
the eggs through the micropyles during fertilization and harmed the eggs, an effect that would
have been minimized in the 0.2 µm filter treated group. It is possible that Acidovorax may have
entered the eggs through the micropyles without harming the egg, unlike the putative pathogen
Flavobacterium sp. B17 and other pathogenic aquatic microbial species present in the stream
water.
Although we cannot definitively determine the mechanism of how fertilization with
Acidovorax sp. F19 lowered egg mortality and yolk resource use, it is worthy to note that there
was a marked difference in colonization success between Acidovorax sp. F19 and
Flavobacterium sp. B17. Our TRFLP revealed that Acidovorax sp. F19 successfully colonized

107

egg surfaces, as opposed to Flavobacterium sp. B17 which did not. 454 pyrosequencing also
detected that significant numbers of Acidovorax spp. were present on the egg surfaces relative to
the number found in the water microbial community, whereas the number of Flavobacterium
was lower on the egg surface relative to the source water microbial community. The rejection of
Flavobacterium populations by the eggs could be mediated by either host innate immunity or
inability of the microbe to adhere to the egg surfaces. The establishment of Acidovorax on the
egg surfaces could be attributed to either its ability to adhere to the surfaces of eggs [36] and/or
evade recognition by the host immune system [10] or its tolerance of the host innate immunity
response [11]. This study thus informs other studies by suggesting that the approach of
characterization of selection for or selection against certain microbes could help identify
potential symbionts and pathogens.
This study also elucidated the mechanisms shaping the egg associated microbial
community and the kinetics of the process of community assembly. The fact that the converged
microbial community at Day 2 post-fertilization was significantly different from that of the
stream water suggests that the egg related local processes (such as host innate immunity) shaped
the community structure. Interestingly, one of the stream water fertilized egg samples shaped the
egg microbial community much faster than others to the point that its community structure at
immediately after fertilization was similar to the community structure found at Day 2. This result
is noteworthy in suggesting that a microbial community can be shaped as a result of egg-related
effects in as quickly as 60 minutes. This finding suggests that there may be a sense of urgency on
the part of the eggs in needing to control the egg associated microbes as rapidly as possible,
thereby highlighting that the onset of host-microbe interaction may be critical for eggs in
determining the trajectory of their life history.

108

Our results with respect to yolk sac area to body area ratio were also interesting. Yolk sac
area alone was slightly larger in eggs fertilized with Acidovorax sp. F19 than that fertilized in
stream water, and body area was slightly smaller in eggs fertilized with Acidovorax than that
fertilized in stream water. Neither of these differences was statistically significant, but a
significant difference was found when combining the two measures in a form of a ratio, thereby
demonstrating the efficacy of this ratio for this analysis. We also note that the measurement of
the yolk sac area alone for assessment of yolk sac resources may not have been a sufficiently
sensitive measure because the yolk sac has a three dimensional structure (height, width, and
depth) that was not fully captured by our measurement (height and width only). The fact that we
also found that body area increased when fertilized in the stream water suggests that embryos
could be encouraged to grow faster when pathogenic microbes colonize the egg surfaces (as in
the stream condition). However, we did not detect a significant difference in hatch timing among
treatments and controls. Alternatively, metabolic activity itself may have been enhanced by the
presence of pathogenic microbes on the egg surfaces, which could have resulted in eggs
consuming yolk sac resources at a faster rate, leading to an increase in body size.
Another interesting finding of this study was that we showed that Acidovorax sp. may be
vertically transmitted from parent to egg, since Acidovorax sp. F19 was closely grouped with one
of the branches within the beta proteobacteria clade found on unfertilized eggs. Furthermore, we
found that unfertilized eggs had similar microbial communities across different females using
TRFLP. This line of evidences suggests that the egg surfaces of unfertilized eggs may not be
sterile, but rather already colonized by certain species of microbes in a non-random manner.
Vertical transmission is an efficient way for hosts to select appropriate symbionts for eggs, since
only effective symbionts can protect eggs and be successfully transmitted to larvae [37].

109

The most important implication for future management of this threatened fish species in
the hatchery environment is that treatment conducted during the first 60 minutes of an egg’s life
can cause significant differences in egg mortality and embryonic resource use. This together with
the fact that egg can already select certain microbes over others within that same brief 60 minute
period suggests that the early interaction between microbes and hosts are important determinants
of the trajectory of a sturgeon’s life. We can now draw on multiple lines of evidence to
recommend Acidovorax sp. F19 as a probiotic that can be used to improve survivorship of Lake
sturgeon eggs reared in hatcheries to assist in recovery of this threatened species in the future.

110

References

111

References

1.

Evans AS (1976) CAUSATION AND DISEASE - HENLE-KOCH POSTULATES
REVISITED. Yale Journal of Biology and Medicine 49: 175-195

2.

Denison RF, Kiers ET (2004) Lifestyle alternatives for rhizobia: mutualism, parasitism,
and forgoing symbiosis. FEMS Microbiology Letters 237: 187-193

3.

Wilson ACC, Ashton PD, Calevro F, Charles H, Colella S, Febvay G, Jander G, Kushlan
PF, Macdonald SJ, Schwartz JF, Thomas GH, Douglas AE (2010) Genomic insight into
the amino acid relations of the pea aphid, Acyrthosiphon pisum, with its symbiotic
bacterium Buchnera aphidicola. Insect Molecular Biology 19: 249-258

4.

Harris RN, Brucker RM, Walke JB, Becker MH, Schwantes CR, Flaherty DC, Lam BA,
Woodhams DC, Briggs CJ, Vredenburg VT, Minbiole KPC (2009) Skin microbes on
frogs prevent morbidity and mortality caused by a lethal skin fungus. ISME J 3: 818-824

5.

McFall-Ngai MJ (2000) Negotiations between animals and bacteria: the [`]diplomacy' of
the squid-vibrio symbiosis. Comparative Biochemistry and Physiology - Part A:
Molecular & Integrative Physiology 126: 471-480

6.

Schramm A, Davidson SK, Dodsworth JA, Drake HL, Stahl DA, Dubilier N (2003)
Acidovorax-like symbionts in the nephridia of earthworms. Environmental Microbiology
5: 804-809

7.

Bright M, Bulgheresi S (2010) A complex journey: transmission of microbial symbionts.
Nature Reviews Microbiology 8: 218-230

8.

Feldhaar H, Gross R (2008) Immune reactions of insects on bacterial pathogens and
mutualists. Microbes and Infection 10: 1082-1088

9.

Gross R, Vavre F, Heddi A, Hurst GDD, Zchori-Fein E, Bourtzis K (2009) Immunity and
symbiosis. Molecular Microbiology 73: 751-759

10.

Siozios S, Sapountzis P, Ioannidis P, Bourtzis K (2008) Wolbachia symbiosis and insect
immune response. Insect Science 15: 89-100

11.

Hooper LV (2009) Do symbiotic bacteria subvert host immunity? Nature Reviews
Microbiology 7: 367-374

12.

Rio RVM, Maltz M, McCormick B, Reiss A, Graf J (2009) Symbiont Succession during
Embryonic Development of the European Medicinal Leech, Hirudo verbana. Appl
Environ Microbiol 75: 6890-6895

112

13.

Ciche TA, Kim K-s, Kaufmann-Daszczuk B, Nguyen KCQ, Hall DH (2008) Cell
Invasion and Matricide during Photorhabdus luminescens Transmission by
Heterorhabditis bacteriophora Nematodes. Appl Environ Microbiol 74: 2275-2287

14.

Hosokawa T, Kikuchi Y, Fukatsu T (2007) How many symbionts are provided by
mothers, acquired by offspring, and needed for successful vertical transmission in an
obligate insect–bacterium mutualism? Molecular Ecology 16: 5316-5325

15.

Mira A, Moran NA (2002) Estimating Population Size and Transmission Bottlenecks in
Maternally Transmitted Endosymbiotic Bacteria. Microbial Ecology 44: 137-143

16.

Lee OO, Chui PY, Wong YH, Pawlik JR, Qian P-Y (2009) Evidence for Vertical
Transmission of Bacterial Symbionts from Adult to Embryo in the Caribbean Sponge
Svenzea zeai. Applied and Environmental Microbiology 75: 6147-6156

17.

Enticknap JJ, Kelly M, Peraud O, Hill RT (2006) Characterization of a Culturable
Alphaproteobacterial Symbiont Common to Many Marine Sponges and Evidence for
Vertical Transmission via Sponge Larvae. Applied and Environmental Microbiology 72:
3724-3732

18.

Usher KM, Kuo J, Fromont J, Sutton DC (2001) Vertical transmission of cyanobacterial
symbionts in the marine sponge Chondrilla australiensis (Demospongiae). Hydrobiologia
461: 9-13

19.

Peterson D, Vecsei P, Jennings C (2007) Ecology and biology of the lake sturgeon: a
synthesis of current knowledge of a threatened North American Acipenseridae. Reviews
in Fish Biology and Fisheries 17: 59-76

20.

Carlson DM (1995) LAKE STURGEON WATERS AND FISHERIES IN NEW-YORKSTATE. Journal of Great Lakes Research 21: 35-41

21.

Williams JE, Johnson JE, Hendrickson DA, Contrerasbalderas S, Williams JD,
Navarromendoza M, McAllister DE, Deacon JE (1989) FISHES OF NORTH-AMERICA
ENDANGERED, THREATENED, OR OF SPECIAL CONCERN - 1989. Fisheries 14:
2-20

22.

Baker EA, Borgeson DJ (1999) Lake Sturgeon abundance and harvest in Black Lake,
Michigan, 1975 -1999. North American Journal of Fisheries Management 19: 1080-1088

23.

Fischer UR, Velimirov B (2000) Comparative study of the abundance of various bacterial
morphotypes in an eutrophic freshwater environment determined by AODC and TEM.
Journal of Microbiological Methods 39: 213-224

24.

Colombo RE, Garvey JE, Wills PS (2007) A guide to the embryonic development of the
shovelnose sturgeon (Scaphirhynchus platorynchus), reared at a constant temperature.
Journal of Applied Ichthyology 23: 402-410

113

25.

R Development Core Team (2009) R: A launguage and environment for statistical
computing. R foundation for Statistical Computing, Vienna, Austria.

26.

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

27.

Marsh TL (1999) Terminal restriction fragment length polymorphism (T-RFLP): An
emerging method for characterizing diversity among homologous populations of
amplification products. Current Opinion in Microbiology 2: 323-327

28.

Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS,
McGarrell DM, Marsh T, Garrity GM, Tiedje JM (2009) The Ribosomal Database
Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Research
37: D141-D145

29.

Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naïve Bayesian Classifier for Rapid
Assignment of rRNA Sequences into the New Bacterial Taxonomy. Applied and
Environmental Microbiology 73: 5261-5267

30.

Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics
Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24: 15961599

31.

Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing
phylogenetic trees. Molecular Biology and Evolution 4: 406-425

32.

Tamura K, Nei M, Kumar S (2004) Prospects for inferring very large phylogenies by
using the neighbor-joining method. Proceedings of the National Academy of Sciences of
the United States of America 101: 11030-11035

33.

Ruby E, Henderson B, McFall-Ngai M (2004) Microbiology - We get by with a little help
from our (little) friends. Science 303: 1305-1307

34.

Dethlefsen L, McFall-Ngai M, Relman DA (2007) An ecological and evolutionary
perspective on human–microbe mutualism and disease. Nature 449: 811-818

35.

Cherr GN, Clark WH (1982) Fine Structure of the Envelope and Micropyles in the Eggs
of the White Sturgeon, Acipenser transmontanus Richardson. Development, Growth &
Differentiation 24: 341-352

36.

Aeckersberg F, Lupp C, Feliciano B, Ruby EG (2001) Vibrio fischeri Outer Membrane
Protein OmpU Plays a Role in Normal Symbiotic Colonization. Journal of Bacteriology
183: 6590-6597

37.

Douglas AE (2008) Conflict, cheats and the persistence of symbioses. New Phytologist
177: 849-858

114

CHAPTER 5: THE RELATIVE IMPORTANCE OF REGIONAL DISPERSAL AND
LOCAL DETERMINISTIC PROCESSES IN SHAPING THE MICROBIAL
COMMUNITY ASSEMBLY ON THE EGG SURFACES OF THE LAKE STURGEON
(ACIPENSER FULVESCENS)

Abstract
Research has shown that both dispersal process and local deterministic processes shape microbial
community assembly, but few studies have been able to analyze the relative importance of these
two processes. We investigated the process of microbial succession on the egg surfaces of the
Lake Sturgeon (Acipenser fulvescens) with a focus on the relative importance of dispersal and
local deterministic processes in shaping the egg microbial community assembly. We
experimentally controlled the rate of dispersal of aquatic microbes onto the egg surface by
manipulating the concentration of water microbes in order to understand the importance of the
rate of dispersal relative to deterministic processes, which in this case was the effect of the egg
micro-environment (e.g. lysozyme, metabolites, and microbe-microbe interactions). Eggs were
fertilized in different water types (stream water, UV-treated stream water, 0.2 µm filtered stream
water) which varied by microbial quantity and community structure, and reared in either the
same water type or one of the other two water types. Eggs were collected at different time points
during incubation. Genomic DNA was extracted from the egg surface and microbial
communities were examined using 16S rDNA-based TRFLP and 454 pyrosequencing. TRFLPderived microbial community data were subsequently analyzed using principal components
analysis (PCA). We hypothesized that dispersal was a dominant process over local deterministic
processes in shaping the egg surface microbial community and the effect of dispersal was

115

dependent on water microbial density. We found that at the rate of dispersal we tested, local
deterministic processes were dominant over dispersal. The egg microbial community assembly
was significantly different from the source water microbial community. However, dispersal also
played a role in explaining the egg surface microbial community. When eggs were reared in
water with a high density of microbes, initial inocula on the egg surface were not important in
dictating the final community composition. However, as the density of water microbes declined,
the final microbial community assemblage on the egg surface became less influenced by the
water in which eggs were reared. In this case, the majority of change accrued in the egg microenvironment was nested within the initial inocula and did not appear to be influenced by
dispersal. These results suggest that local deterministic processes are dominant over dispersal,
and the effect of dispersal on the egg microbial community was dependent on the concentration
of aquatic microbes present during microbial succession.

116

Introduction
Microbial communities vary spatially and temporally. Mechanisms shaping community
assemblage at a given place and time remain fundamental questions in ecology [1]. The study of
microbial communities has flourished over the last 15 years, partly due to the development of
analytical techniques that have allowed us to expose unculturable portions of such communities
[2-3]. Microbial community assemblages have been studied in a variety of environments ranging
from deep marine sediments [4], marine water [5], lakes [6-7], soils [8-9], plants [10-11], the
animal gut [12], and the human gut [13]. However, the underlying mechanisms shaping
microbial community assemblage at a given place and time remain largely unknown, as does the
role of microbes in the greater ecological system.
In the realm of ecology, there are two competing theories that seek to explain the
processes shaping community assemblage: niche theory [14] and neutral theory [15]. Niche
theory assumes that organisms are selected by deterministic processes for functional traits that
allow them to exist in certain environments. According to this theory, coexistence among
different species at a given location is explained by heterogeneity of the local environment (local
process) [16]. On the other hand, neutral theory assumes that organisms are functionally neutral
in a given environment. Stochastic processes such as dispersal explain presence or absence of
organisms in a given environment. In this context, dispersal refers to movement of organisms
such as plant seeds, insects, and microbes away from an existing population, a process which is
mediated by passive forces such as wind and currents. According to this theory, community
assemblage at a given location is shaped by dispersal from neighboring habitats (a regional
process) [15].

117

Several studies have attempted to elucidate the relative importance of these two processes
in shaping community assemblage using microbial systems [17-18]. Dumbrell and colleagues
studied the relative importance of soil pH and microbial dispersal on soil microbial community
structure and found that pH was a more dominant factor over dispersal [17]. Van der Gucht and
colleagues studied the relative importance of local processes and regional processes (dispersal) in
explaining eleven lake microbial community assemblages, and found that local processes were
more dominant [18]. However, these studies are missing one of the key components that may
explain the observed microbial community assemblage in a given environment- the history of the
community [19].
History can effectively be understood via the study of succession. Succession is, by
definition, the colonization of an open space and subsequent sequential changes in species
composition. Succession is also a special type of community assembly in which the entire
process of community development can be observed. Microbial succession has been recently
studied in various host animals [19-21], host plants [10], and natural [22-24] and artificial [25-27]
environments. These previous studies found that microbial succession is a complex process
affected by a number of factors [10, 19, 22]. Key factors include initial dispersal of microbes
from the neighboring communities followed by colonization in or attachment to the open space,
and subsequent species sorting via local deterministic processes, while microbes are
continuously immigrating from neighboring spaces [28].
Microbes serve as good model systems for studying these processes. This is partly
because microbial generation time is relatively short, which allows us to observe the process of
adaptation. Microbes also exhibit high dispersal rates, since their small size allows them to
disperse freely without geographic barriers [29]. In addition, microbes inhabit various

118

environments, which allow us to observe the effect of environmental gradients on shaping
community assemblage.
One unique host for which these processes have not yet been studied is the egg surface of
the threatened fish species Lake sturgeon (Acipenser fulvescens). Microbial succession on the
Lake sturgeon egg surface is likely a complex process. Eggs are fertilized in a stream as soon as
male and female adults release gametes. The fertilized eggs develop stickiness [30] so that they
adhere to benthic substrates such as gravel and sand [31]. During this fertilization process,
microbes that are drifting in a stream collide with eggs and adhere to the sticky egg surfaces.
This process is a stochastic process, since the water microbial community varies temporally and
spatially [24, 32-33], and eggs in the stream have no control over their movements. After the
initial stochastic collision, the microbial community on egg surfaces is possibly selected by local
deterministic processes including adhesion [34-35], antimicrobial activities of eggs [36-37],
chemicals that eggs excrete during embryogenesis [38-39], interspecific competition among
microbial species [40], while microbes in stream water continuously collide with egg surfaces
via passive dispersal mediated by water flow.
In this study, we investigated the relative importance of local deterministic processes and
dispersal in shaping microbial community assemblage on the Lake Sturgeon egg surface during
microbial succession. The novel contribution of this study is that we can quantitatively assess the
relative importance of deterministic and stochastic processes in explaining microbial community
assemblage on the egg surface by manipulating the quality, quantity, timing, and duration of
microbial dispersal, unlike a field observational study where there is no such control. We
hypothesized that dispersal was a dominant process over local deterministic processes in shaping
the egg surface microbial community and the effect of dispersal was dependent on water

119

microbial density. This study has significant implications for understanding mechanisms
governing microbial succession and also for the management of the threatened Lake Sturgeon,
which is susceptible to high egg mortality as a result of colonization of microbes.

Methods
Study site
This experiment was conducted at a Lake Sturgeon streamside hatchery located on the
Upper Black River in Michigan during May 2011 in the midst of the Lake Sturgeon spawning
season. Incoming river water was filtered using a sock filter system (100 psi) to remove large
particulate matter before being gravity-fed to the hatchery system. Gametes used in this study
were collected from spawning adults in the Upper Black Lake and fertilized and reared in the
hatchery under different experimental conditions.

Experimental design and sampling
Three different water types (stream water, UV treated stream water, and 0.2 µm filtered
stream water) which varied by microbial quantity and community composition were used in this
experiment. UV treated water was created using a water treatment system which consisted of a
50µm filter cartridge followed by a UV lamp (Emperor aquatic, Inc). The UV treated water was
circulated in the system for 24 hours before being fed to egg samples. The 0.2 µm filtered water
was obtained by filtering the hatchery water using a series of filer cartridges- 50 µm, 20 µm, 10
µm, 5um, 1.0 µm (Pentair Ltd) and 0.2 µm (Spectrapure Inc). The 0.2 µm filtered water was
stored in a reservoir and filtered again using a series of 1.0um and 0.2 µm filter cartridges before
being fed to egg samples.

120

Approximately, 100 eggs were fertilized in one of the three water types (stream water,
UV treated stream water, or 0.2 µm filtered stream water), and continually reared in the same
o

water type for 6 hours at 18-19 C. Six hours after fertilization, eggs were either continually
o

reared in the same water or reared in different water types at 18-19 C. Thus, this experiment
consisted of a total of 9 treatments with different fertilization/rearing combinations (Figure 5.1).
Rearing water was re-circulated for stream water and UV treated water, while 0.2 µm filtered
water was not re-circulated in order to prevent microbial contamination from eggs. Flow rate of
this experiment was maintained at 5 L per minute for all water types. The experiment was
replicated using 6 families. Egg samples were collected from each treatment at four different
time points (AF, Day 1, Day 3, and Day 5) during incubation. Water samples were collected for
each water type at the same time points that the eggs were collected. For each water sampling
event, 100 mL of water was collected and filtered with a 0.22 µm filter membrane (manufacture)
o

and stored in 20 mL ethanol at 4 C.

Direct microscopic counts
One mL of each of the water microbial samples (stream water, UV treated water, and 0.2
µm filtered water at four different time points) that were stored in 80% ethanol was placed in a
1.5 mL eppendorf tube and centrifuged at 11,000 rpm for 3 minutes. The cell pellets were then
re-suspended in 500 uL water. 10 uL of 0.38% Gram Crystal Violet (DIFCO) was added into the
cell suspension and was left to sit for 3 minutes to stain the cells. The cell suspension was
centrifuged at 11,000 rpm for 3 minutes. The cell pellets were re-suspended in either 100 uL

121

Figure 5.1. Schematic diagram of experimental design to manipulate initial inocula and
subsequent rate of dispersal.
Eggs were fertilized in 3 different water types (stream water, UV treated stream water, or 0.2 µm
filtered stream water). 6 hours after fertilization, eggs were transferred and reared in the same or
different water types. A total of 9 different fertilization/rearing combinations were maintained.
Six families have been tested in this manner.
122

water (for stream water samples) or 10 uL water (for UV treated and 0.2 µm filtered water
samples). Three uL of the cell suspension was placed on a Petroff-Hausser chamber. The number
of cells on each of 16 out of 25 middle-sized grids was counted using a light microscope with a
total magnification of 1000X. The number of cells per mL of original water sample was
calculated based on the volume of the middle sized grids and the concentration factor of the
samples.

Extraction of DNA
A total of 12 water community samples (3 water types, 4 time points) and 40 egg
microbial community samples (2 families, 5 out of 9 treatments, 4 time points) were processed
for DNA extraction. For water samples, each water sample was vortexed for 10 minutes with
maximum vibration, and 10 mL of the ethanol solution containing suspended bacterial cells were
o

taken into a 15 mL corex tube and centrifuged at 10,000 rpm for 30 minutes at 6 C. Cell pellets
from the centrifugation were subjected to DNA extraction using the Power Soil

TM

Kit (MO BIO

Laboratories Inc., CA). For egg samples, genomic DNA was extracted from the surface of 8 eggs
per sample using the Power Soil

TM

Kit according to the manufacture’s protocol.

454 pyrosequencing analysis
To characterize both egg associated microbial communities and water microbial
communities in each water type, the extracted DNA samples were subjected to 454
pyrosequencing. The V3-V5 region of the 16S rRNA gene of the extracted DNA (see section
above) was sequenced using 454 GS FLX titanium platform (454 Life Science, Branford, CT) at

123

a research facility in Baylor, Texas. Raw sequence reads were processed using Ribosomal
Database Project (RDP) pipeline [41] to sort the data by tag sequence, to trim tag and primer
sequences, and to filter out low quality sequences with a minimum quality score of 20
(probability threshold of 0.01) and minimum read length of 300bp. The taxonomy of the filtered
reads was assigned using RDP Classifier at a bootstrap threshold of 80% [42]. The resultant
microbial communities of the samples were compared at both the phylum and genus level.

Community analysis using TRFLP
16S rRNA gene based TRFLP was performed to characterize microbial community
structure [43-44]. The detailed PCR amplification procedures for TRFLP were described in
Chapter 2. The purified PCR products were subjected to enzyme digestion with HhaI (Gibco
BRL). Two technical replicates of each of the digested DNA samples were sent to Michigan
State University’s sequencing facility and the DNA fragments were separated on an ABI 3100
Genetic Analyzer automated sequencer (Applied Biosystems Instruments, Foster City, CA) in
GeneScan mode. The sizes and abundance (peak height) of the terminal restriction fragments
(TRFs) were calculated using GeneScan 3.7. Each terminal fragment corresponds to a phylotype,
and peak height indicates relative abundance of a phylotype. In order to align TRFs across egg
samples from different treatments, the TRFLP profiles were processed with T-Align software
(http://inismor.ucd.ie/~talign/index.html) and the output of T-Align was used for the microbial
community analysis. Principal component analysis (PCA) was performed on the TRFLP data in
order to elucidate underlying patterns across samples. PCA was conducted using the “prcomp”
function of the R software version 2.10.0 [45].

124

Quantitative PCR analysis
To assess the effect of the various treatments on the egg associated microbial quantity,
microbial load of egg samples from different treatments were determined by performing
quantitative PCR (qPCR) with SYBR green. 2 families were included in the analysis. The qPCR
was performed using the same protocol described in Chapter 3. A standard curve for the
relationship between 16S rRNA gene copy number and cycle threshold (Ct) values was
constructed using a series of dilutions of the bacterial genomic DNA Flavobacterium johnsoniae
ATCC 17061 that is known to have six 16S rRNA gene copies in its genome. The quantity of the
16S rRNA gene copy of each sample was determined by substituting the Ct value of one of the
sample dilutions into the equation of the standard curve and multiplying it by the dilution factor.

Larval size analysis
We were also interested in examining the effect of rearing environments on larval size of
eggs at hatch. Immediately after the hatching of eggs, larvae were anesthetized using MS-222.
Twenty individuals per treatment per family were photographed with a ruler as a size standard.
Larvae from 3 (fertilized/reared in stream water, UV treated, and 0.2 µm filtered) out of 9
treatments were analyzed for the larval size. 6 families were included in the analysis. For each
larvae, the total length, total body area, and yolk sac area were determined from the images using
ImageJ software. Yolk sac area was chosen for measurement because this allowed us to examine
the effect of treatment on the allocation of yolk resources during embryogenesis. The effect of
fertilization/incubation in 0.2 µm filtered water on the larvae size was assessed using a general
linear model, and the statistical significance of the treatment effect on larvae size was confirmed
after accounting for family effect as a random variable using a linear mixed effect model. The

125

general linear model was performed using the “lm” function and the linear mixed effect model
was performed using “lme” function in R version 2.10.0 [45].

Assessment of egg mortality
The death of an egg was defined as the arrest of embryonic development. The arrest of
embryonic development was determined by visual observation of developmental stages of
embryos relative to a reference [46]. The number of dead eggs was recorded for each treatment
and control on a daily basis, and all dead eggs were removed from the incubation tray at
detection. The number of successful hatches for each treatment and control was also recorded.
The cumulative egg mortality for each treatment and control was calculated as follows: Egg
mortality = total number of dead eggs / (total number of dead eggs + total number of hatches).
The cumulative egg mortality was compared across treatments using a box plot and the effect of
treatments on egg mortality was assessed using a general linear model using the “lm” function in
the R software version 2.10.0 [45].

Results
Source water microbial quantity and compositions
Microbial concentrations in each water type were estimated using direct microscopic
counts with Petroff-Hauser counting chamber. The average concentrations of microbes for
stream water, UV treated water, and 0.2 µm filtered water were found to be 10
3.29

10

5.28

, 10

4.22

,

16S rRNA gene copies per 1 mL, respectively (Figure 5.2). The concentration was the

highest in stream water, the second highest in UV treated water, and the lowest in 0.2 µm filtered
water. There was approximately one order of magnitude difference between stream and UV
126

treated water, and another one order of magnitude difference between the UV treated and 0.2 µm
filtered water (Figure 5.2).

Figure 5.2. Direct microscopic counts for microbial density in each water type.

Microbial composition of each water type was characterized using 454 pyrosequencing at
both the phylum/class level (Table 5.1). A significant difference in water microbial community
structure among the three water types was detected at the phylum/class level. The stream water
microbial community was dominated by beta-Proteobacterium, Bactroidetes, and Actinobacteria.
UV treatment decreased the relative abundance of Bacteroidetes and beta-Proteobaccteria, and
increased the relative abundance of Actinobacteria, Deinococcus-Thermus, and Firmicutes. The
relative abundance of alpha-Proteobacteria and Chloroflexi increased after treatment with 0.2 µm

127

filtration. The microbial community composition in each water type was also analyzed at the
genus level and the results are provided in Table 5.2 for reference.

Microbial community analysis using TRFLP
The effect of water type on the egg surface microbial communities was examined using
PCA with TRFLP data. Principal component analysis showed that the microbial communities on
the egg surfaces were clustered by water types in which they were fertilized and reared (Figure
5.3). Within each water type, there was a directional temporal trend along with egg
developmental stages (Figure 5.3). Temporal patterns were more explicitly observed for eggs
fertilized in stream water and 0.2 µm filtered water.
We further tested whether the effect of dispersal is dependent on the concentration of
aquatic microbes. PCA showed that when eggs were fertilized in 0.2 µm filtered water and
subsequently reared in stream water, the egg surface microbial community assembly converged
with the community on eggs fertilized and reared in stream water (arrow in Figure 5.4). However,
when eggs were fertilized in stream water and subsequently reared in 0.2 µm filtered water, the
egg surface microbial community assembly did not converge with the community on eggs
fertilized and reared in 0.2 µm filtered water (Figure 5.4), but rather was clustered with the
community on eggs fertilized and reared in stream water.

128

Table 5.1. Comparison of microbial communities in different water types using 454
pyrosequencing analyzed at the phylum/class level.

Phylum/Class
Betaproteobacteria
Bacteroidetes
Actinobacteria
α-proteobacteria
Firmicutes
Cyanobacteria
γ-proteobacteria
OD1
δ-proteobacteria
un-proteobacteria
Verrucomicrobia
Acidobacteria
Chloroflexi
TM7
ε-proteobacteria
Gemmatimonadetes
Planctomycetes
Nitrospira
Chlamydiae
Fusobacteria
Armatimonadetes
OP11
Unclassified
phylum

Stream
AF
(%)
46.3
29.6
7.4
4.4
1.8
1.3
1.2
1.1
0.8
0.8
0.7
0.2
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.0
0.0
0.0

UV
AF
Phylum/Class
(%)
Actinobacteria
23.8
α-proteobacteria
17.9
Deinococcus-Thermus
15.9
β-proteobacteria
11.9
Firmicutes
7.3
Chloroflexi
7.3
Bacteroidetes
5.3
γ-proteobacteria
3.3
Cyanobacteria
0.7
Chlorobi
0.7
Unclassified phylum
6.0

Phylum/Class
α-proteobacteria
Chloroflexi
β-proteobacteria
Bacteroidetes
γ-proteobacteria
Actinobacteria
Deltaproteobacteria
Firmicutes
TM7
Armatimonadetes
Acidobacteria
Unclassified phylum

0.2
AF
(%)
35.8
17.9
12.3
9.9
4.3
4.3
3.1
3.1
0.6
0.6
0.6
7.4

3.6

Samples were collected from each water type at 6 hours after fertilization. Numbers in the table
indicate the relative abundance of each phylum in each water community. Total reads for each
sample were 6434, 151, and 162 for stream AF, UV AF, and 0.2 µm AF, respectively. “AF”
stands for after fertilization.

129

Table 5.2. Comparison of microbial communities in different water types using 454
pyrosequencing analyzed at the genus level.
Genus
Limnohabitans
Flavobacterium
Polynucleobacter
OD1_incertae_sedis
Fluviicola
Sphaerotilus
Acidovorax
Methylophilus
Arcicella
Albidiferax
Bacillariophyta
Sphingomonas
Aquabacterium
Massilia
Clostridium sensu
Opitutus
Brevundimonas
Sediminibacterium
Algoriphagus
Leptothrix
Polaromonas
Rhodobacter
Caulobacter
Other classified
Unclassified

Stream
AF (%)
22.1
13.2
3.2
1.1
1.0
0.9
0.9
0.9
0.9
0.8
0.8
0.5
0.4
0.4
0.4
0.4
0.4
0.4
0.3
0.3
0.3
0.3
0.2
5.6
44.2

UV AF
0.2 AF
Genus
(%)
Genus
(%)
Microbacterium
18.5 Asticcacaulis
16.7
Deinococcus
15.9 Cytophaga
6.2
Flavobacterium
4.6 Brevundimonas
6.2
Brevundimonas
4.6 Caulobacter
5.6
Caulobacter
3.3 Delftia
1.9
Novosphingobium
3.3 Limnohabitans
1.9
Sphaerotilus
2.0 Pseudomonas
1.9
Schlegelella
2.0 Sphaerotilus
1.2
Asticcacaulis
2.0 Cellvibrio
1.2
Pseudomonas
1.3 Bacillus
1.2
Bacillus
1.3 Thermopolyspora
0.6
Sphaerobacter
1.3 TM7_incertae_sedis
0.6
Aquabacterium
0.7 Fluviicola
0.6
Hydrogenophaga
0.7 Leptothrix
0.6
Massilia
0.7 Polynucleobacter
0.6
Naxibacter
0.7 Acidovorax
0.6
Aquamicrobium
0.7 Hydrogenophaga
0.6
Methylobacterium
0.7 Pelomonas
0.6
Sphingobium
0.7 Methylophilus
0.6
Rhodobacter
0.7 Bosea
0.6
Acinetobacter
0.7 Rhizobium
0.6
Streptophyta
0.7 Pedomicrobium
0.6
Ignavibacterium
0.7 Phenylobacterium
0.6
Paenibacillus
0.7 Other classified
4.3
Unclassified
31.8 Unclassified
43.8

Samples were collected from each water type at 6 hours after fertilization. Numbers in the table
indicate the relative abundance of each genus in each water community. Total reads for each
sample were 6434, 151, and 162 for stream AF, UV AF, and 0.2 µm AF, respectively. “AF”
stands for after fertilization.

130

Figure 5.3. The effect of water type (aquatic microbial community) on the egg surface
microbial communities as measured using T-RFLP.
PCA shows that eggs that were fertilized and reared in stream water had different egg surface
microbial communities from those reared in either UV-treated or 0.2 µm filtered water. “AF”,
“D1”, “D3”, and “D5” stand for 6 hours after fertilization, Day 1, Day 3, and Day 5 postfertilization, respectively.

131

Figure 5.4. PCA plot showing that dispersal was dependent on water microbial density.
When eggs were fertilized in 0.2 µm filtered water and subsequently reared in straight stream
water, the egg surface microbial community assembly converged with those on eggs fertilized
and reared in stream water. However, when eggs were fertilized in stream water and
subsequently reared in 0.2 µm filtered water, the egg surface microbial community assembly did
not converge with those of eggs fertilized and reared in 0.2 µm filtered water. “AF”, “D1”, “D3”,
and “D5” stand for 6 hours after fertilization, Day 1, Day 3, and Day 5 post fertilization,
respectively.

132

We compared water microbial communities and the egg microbial communities using
PCA with TRFLP data. Our hypothesis was that if the dispersal is dominant over deterministic
processes, the microbial community assembly of the egg surfaces is similar to that of the source
water. PCA revealed that the egg associated microbial communities were not similar to the
source aquatic microbial communities (Figure 5.5). Egg surface microbial communities were
clustered together and were separated from water microbial communities in the PCA plane.
Although the egg microbial communities were clustered tightly (red circle in Figure 5.5), there
was an observable effect of aquatic microbial community on the egg microbial community as
evident by the clustering by distinct water type within this egg cluster. This effect of the water
microbial community on the egg surfaces is small compared to the local egg effect, which
separated egg microbial communities from the source water microbial communities. There was
also a temporal trend detected in this PCA plane (Figure 5.5). Microbial communities on eggs
collected 6 hours after fertilization were closer to the source water microbial communities, and
the egg microbial communities diverged from the source water microbial communities as eggs
developed.

133

Figure 5.5. PCA analysis on the relative importance of dispersal and deterministic
processes in community assembly as measured with T-RFLP.
PC1 and PC2 separated egg surface microbial communities from water microbial communities.
The red circle indicates a cluster of egg microbial communities. “AF” denotes samples collected
6 hours after fertilization and “D5” denotes samples collected at Day 5 post-fertilization.

134

454 pyrosequencing analysis
We performed 454 pyrosequencing for both stream water fertilized egg samples and the
source stream water sample to answer a question about what microbes were selected for or
against by eggs. Pyrosequencing analysis revealed a difference in the microbial community
assembly between the egg surfaces and the source water at the phylum level (Table 5.3). Three
phyla including Bacteroidetes, Actinobacteria, and Firmicutes were found to be less abundant on
the egg surfaces than the source water. Genus level analysis revealed that genera including
Albidiferax, Roseateles, Hydrogenophaga, Sphaerotilus, Aquabacterium, Caulobacter,
Pseudomonas, and others were found to be more abundant on egg surfaces than in the source
water, while genera including Polynucleobacter, Limnohabitans, and Flavobacterium were
found to be less abundant on the egg surfaces relative to the source water (Table 5.4).

135

Table 5.3. Comparison between the egg associated microbial communities and source water
microbial communities at 6 hours after fertilization using 454 pyrosequencing analyzed at
the phylum level.
Phylum
*QB Egg SS AF
RC Egg SS AF
Stream Water AF
Proteobacteria
3629
4254
3449
Firmicutes
1
2
116
Bacteroidetes
710
768
1905
Chloroflexi
1
2
7
Nitrospira
0
1
4
Actinobacteria
60
114
477
TM7
0
7
6
Acidobacteria
17
3
16
Fusobacteria
0
0
1
Verrucomicrobia
25
40
48
Gemmatimonadetes
1
2
5
Planctomycetes
1
0
5
Armatimonadetes
2
8
1
OD1
6
6
70
Chlamydiae
0
0
4
Cyanobacteria/Chloroplast
109
181
86
Deinococcus-Thermus
2
1
0
OP11
0
1
1
Unclassified phylum
320
175
233
Total reads
4884
5565
6434
*“QB” and “RC” are family codes used for this experiment. “SS” denotes a treatment in which
eggs were fertilized in stream water and reared in stream water. “AF” stands for 6 hours after
fertilization.

136

Table 5.4. Comparison of the egg associated microbial communities and source water
microbial communities at 6 hours after fertilization using 454 pyrosequencing analyzed at
the genus level.
Genus
Albidiferax
Roseateles
Hydrogenophaga
Sphaerotilus
Aquabacterium
Rubrivivax
Ideonella
Duganella
Novosphingobium
Caulobacter
Rhodobacter
Pseudomonas
Ferruginibacter
Flectobacillus
Polynucleobacter
Limnohabitans
Flavobacterium
Sediminibacterium
Other classified
Total unclassified
Total reads

*QB Egg
SS-AF
199
14
59
850
117
10
366
44
71
49
147
24
16
172
11
109
177
2
510
1937
4884

RC Egg
SS-AF
201
13
57
937
113
15
397
33
217
108
231
48
11
81
28
101
245
1
826
1902
5565

Stream
Water AF
53
0
12
61
27
0
8
4
5
16
17
2
1
2
204
1419
850
24
884
2845
6434

Abundance
on eggs
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Higher
Lower
Lower
Lower
Lower

* The 18 genera that exhibited a difference in abundance between the egg surface and source
water are shown. “QB” and “RC” are family codes used for this experiment. “SS” denotes a
treatment in which eggs were fertilized in stream water and reared in stream water. “AF” stands
for 6 hours after fertilization.

137

454 pyrosequencing was also performed to examine the effect of water types on the egg
microbial community assembly and the temporal trend within each treatment group (Table 5.5).
We found a trend in which some genera were more closely associated with eggs fertilized and
reared in certain water types (Table 5.5). Genera including Acidovorax, Methyloversatilis,
Shinella, Bosea, Asticcacaulis, Caulobacter, Pseudomonas, Cellvibrio, Bdellovibrio, and
Bacteriovorax were strongly associated with eggs fertilized and reared in 0.2 µm filtered water,
while genera including Naxibacter, Ferribacterium, Brevundimonas, and Deinococcus were
strongly associated with eggs fertilized and reared in UV treated water (Table 5.5). Some genera
were strongly associated with eggs fertilized and reared in both 0.2 µm filtered water and UV
treated water, which were genus Roseateles and Hydrogenophaga. The abundance of the egg
associated microbes, including genus Hydrogenophaga, Methyloversatilis, Shinella,
Methylophilus, Bosea, Bdellovibrio, and Bacteriovorax increased along with the egg
development. The abundance of the egg associated microbes including genus Ideonella,
Asticcacaulis, Caulobacter decreased with egg development. The temporal trend was consistent
regardless of rearing water type.

138

Table 5.5. Water treatment effect on egg microbial community plus temporal trend analyzed using 454 pyrosequencing
at genus level (RC family data).

“AF”, “D1”, “D3”, and “D5” stand for 6 hours after fertilization, Day1, 3, and 5 post fertilization, respectively. “Water Type”
implies water type affiliation that a genus exhibited on the eggs. “Peak Time” implies the temporal affiliation that a genus
exhibited on eggs.

139

Table 5.5 (cont’d)

140

Quantitative PCR analysis
The microbial quantity on the surfaces of eggs fertilized and reared in stream water
6

increased from 10 to 10

7.5

16S rRNA gene copies per egg (Figure 5.6). Eggs fertilized in UV

treated and 0.2 µm water had lower microbial quantities compared to that fertilized in stream
water throughout incubation, except at Day 3 when microbial quantity on eggs fertilized in all
three water types were similar to each other. The microbial load on the egg surfaces fertilized
and reared in UV treated water increased from 10

5.4

7

to 10 16S rRNA gene copies per egg

during embryogenesis, while eggs fertilized and reared in 0.2 µm filtered increased from 10
6.5

10

5.5

to

16S rRNA gene copies per egg during embryogenesis (Figure 5.7). The surfaces of eggs

fertilized and reared with 0.2 µm filtered water had a slightly greater number of microbes at the
beginning of the incubation period relative to those fertilized and reared in UV filtered water,
although one of the two replicates had an almost identical quantity in the two treatments.
We were interested in investigating how microbial quantity of the eggs fertilized in
stream water and reared in 0.2 µm filtered water changed over time, since we observed that the
egg microbial community fertilized in stream water and reared in 0.2 µm thereafter did not
converge with the f eggs that were fertilized in stream water and reared in 0.2 µm filtered water
followmicrobial community observed on eggs fertilized and reared in 0.2 µm. Microbial quantity
oed the trend of eggs fertilized and reared in stream water, which was noticeably different from
the trend we observed for eggs fertilized and reared in 0.2 µm filtered water (Figure 5.7).

141

Figure 5.6. Quantification of the egg associated microbes across different treatments using
qPCR.
“SS”, “UU”, and “OO” stand for eggs fertilized and reared in stream water, UV treated water,
and 0.2 µm filtered water, respectively.

142

Figure 5.7. The effect of transfer from stream water to 0.2 µm filtered water on the egg
surface microbial quantity.
“AF”, “D1”, “D3”, and “D5” stand for 6 hours after fertilization, Day 1, Day 3, and Day 5 postfertilization, respectively.

143

Larvae size analysis
The effect of rearing environments on yolk sac area was examined. Larvae that were
fertilized and reared in 0.2 µm filtered water had significantly larger yolk sac area (8.35 ± 0.84
2

2

mm ) than those that were fertilized and reared in stream water (7.98 ± 0.65 mm ) (t 315=3.485,
p<0.001, Figure 5.8). There was also a statistically significant difference in yolk sac area
2

between larvae fertilized and reared in UV treated water (8.22 ± 0.78 mm ) and those fertilized
2

and reared in stream water (7.98 ± 0.65 mm ) (t 315=2.247, p=0.03, Figure 8). The effect of
water type on yolk sac area was significant after accounting for the family effect on yolk sac area
with a family as a random variable using the mixed effect model (data not shown).

Egg mortality
The effect of rearing environments on the egg mortality was studied. Eggs that were
exposed to stream water for at least one time point during embryogenesis had higher egg
mortality than eggs exposed to all other treatments (Figure 5.9). Both eggs fertilized and reared
in 0.2 µm filtered water and eggs fertilized in UV treated water and reared in 0.2 µm filtered
water had the lowest egg mortality (33%) of all treatments (Figure 5.9). Eggs fertilized and
reared in stream water had the highest egg mortality (58%) (Figure 5.9). The difference in the
egg mortality between eggs fertilized and reared in 0.2 µm filtered water and fertilized and
reared in stream water was statistically significantly different (t43 = 1.76, p=0.086).

144

Figure 5.8. The effect of rearing water type on yolk sac resource uses.
“SS”, “UU”, and “OO” stand for eggs fertilized and reared in stream water, UV treated water,
2

and 0.2 µm filtered water, respectively. The unit of yolk sac area is mm .

145

Figure 5.9. A box plot showing the effect of fertilization and rearing environment on egg
mortality.
S stands for stream, U stands for UV treated, and O stands for 0.2 µm filtered. The first letter
represents the treatment for fertilization and the second letter represents the treatment for rearing
environment.

146

Discussion
In this study, we demonstrated that both egg-related local deterministic processes and the
dispersal of aquatic microbes onto egg surfaces are important processes in explaining Lake
Sturgeon egg surface microbial communities. At the rate of dispersal we tested, the local
deterministic processes appeared to be a dominant process over dispersal, which did not meet my
hypothesis. However, dispersal was also an important process in shaping the egg surface
microbial communities, and its effect on microbial communities appeared dependent on aquatic
microbe density, as we hypothesized.
Our results revealed that 6 hours after fertilization, the egg microbial community
assembly was already significantly different from the source water microbial community. This
fact suggests that the egg-related deterministic processes acted quickly to shape the community.
We believe that this rapid process was mediated by host innate immunity, specifically immunity
that was maternally provisioned [36, 47-49]. Phylum Actinobacteria, one of the dominant phyla
in water microbial communities, had low representation on the egg surfaces. This is likely
because the host innate immunity (including maternally provisioned lysozyme) acted on the
peptidoglycan layer of this gram positive phylum. Another possible explanation is that the
difference in the community structure between eggs and source water was mediated by the
ability of aquatic microbes to adhere to the egg surfaces [34-35], since not all microbes adhered
to the egg surfaces with equal affinity. Sphaerotilus spp., one of the dominant genera found on
the egg surfaces in this study, is known to possess a sheath structure that facilitates adherence to
solid surfaces [50].
The temporal trend in microbial community succession that we observed was also driven
by deterministic processes. The direction of the community shift observed on the egg surfaces

147

showed a gradual divergence from the source water microbial community. Furthermore, the egg
surface microbial community at Day 5 had significantly diverged from the source water
microbial communities, specifically water samples collected at the same time point. This
suggests that the changes we observed on the egg surface microbial communities were not
derived from the source water microbial communities. In other words, continuous dispersal from
the source water did not seem to affect the development of microbial communities on the egg
surfaces. The local processes that caused the shift in microbial communities could have been
changes in metabolites on the egg surfaces from urea to ammonia [39, 51-52], microbe-microbe
interactions (see Chapter 6), biofilm formation (see Chapter 6), or changes in lysozyme type
from maternally provisioned to egg secreted [53-54]. Hydrogenophaga spp., which are known to
be dominant in stream biofilm [55], were found to increase on the egg surfaces in this study.
We believe that the local deterministic processes consisted of two distinct processes. The
first process was one that acted quickly within 6 hours post fertilization. This process is
representative of factors related to host innate immunity and host surface chemistry that select
for attachment of certain microbes and is the most dominant process of all in explaining the egg
surface microbial community assembly. The second process had more subtle effect on the egg
surface microbial community, but gradually shaped the community over time, a process which is
represented by changes in metabolites and microbe-microbe interactions.
We hypothesized that dispersal of aquatic microbes from water column onto the egg
surfaces is the most dominant factor and the egg microbial communities converge with the
source water microbial community. However, at the rate of dispersal we tested in this study, the
hypothesis did not hold true. Instead, egg microbial communities diverged from the source water
microbial communities over time. However, dispersal did play some role in shaping the

148

microbial communities because egg microbial communities fertilized and reared in different
water types had different egg microbial community assemblages. One significant finding was
that the effect of dispersal on egg microbial community assembly was dependent on the
concentration of water microbes. This is evident when considering the fact that initial inocula on
the egg surface were not important in explaining the final community composition when eggs
were reared in water with a high density of microbes. However, as the density of water microbes
declined, the majority of change accrued in the egg micro-environment was nested within the
initial inocula and did not appear to be influenced by dispersal.
Dispersal did not play significant role in explaining microbial quantity on the egg
surfaces, especially when eggs were transferred from stream water to 0.2 µm filtered water.
Microbial quantity continued to grow even after eggs were transferred to aquatic environment
with a low level of microbial concentration. The microbial community composition that
developed after the transfer from stream water to 0.2 µm filtered water was not similar to that of
fertilized and reared in 0.2 µm filtered water, suggesting that the additional microbial load on the
egg surface did not come from the 0.2 µm filtered water. These lines of evidence suggest that
microbial quantity cannot be directly explained by dispersal and is instead likely attributed to
microbial growth among the initial colonizers.
Eggs fertilized and reared with 0.2 µm filtered water had a slightly greater number of
microbes at the beginning of incubation relative to those fertilized and reared in UV filtered
water, although one of the two replicates showed the almost identical quantity between the two.
This difference in microbial quantity on the egg surface could be due to the fact that the
compositions of microbes in 0.2 µm filtered water were not selected against by eggs, in other
words they successfully colonized eggs. We found that genus Acidovorax, which can effectively

149

colonize the egg surfaces as described in Chapter 4, became dominant at Day 1 on egg surfaces
fertilized and reared in 0.2 µm filtered water. This line of evidence suggests that not only the
quantity of aquatic microbes, but also composition of the aquatic microbial community may be
important in explaining the effect of aquatic microbes on the egg surface microbial quantity.
Our study also demonstrated that history may matter in explaining the future
development of a microbial community. Several lines of evidence suggest that the effect of
initial colonization on development of subsequent microbial communities was likely dependent
on both available space on eggs and/or the concentration of the source water microbes. When
eggs were fertilized in 0.2 µm filtered water for 6 hours and were transferred and reared in the
stream water with high microbial load, the initial microbial community structure was masked and
converged with the community of stream fertilized and reared eggs. However, when the eggs
fertilized in stream water (with high microbial load) for 6 hours were transferred and reared in
0.2 µm filtered water, the egg associated microbial community did not converge with that
fertilized and reared in 0.2 µm filtered water. This suggested that either available space on egg
surface and/or the concentration of the source water microbes plays a role in determining the
subsequent microbial community assembly. This suggests that the history of community
composition may be important particularly when a large impact (large microbial load) occurred
in the initial stage and/or smaller impacts (small microbial dispersal) occurred in the later stages.
Although the effect of dispersal in shaping the microbial community structure was small,
the minor differences in community structure affected the life history traits of the host, including
egg mortality and yolk sac area at hatch. We identified genera that specifically associated with
eggs that had a lower egg mortality and lower yolk sac resource use. We believe that genera
associated with eggs fertilized and reared in 0.2 µm filtered water and/or UV treated water

150

(described in Table 3) are good candidates for probiotic treatment for the fish eggs. In fact,
genera Acidovorax and Pseudomonas which we identified as putative symbionts for the sturgeon
eggs in other chapters (Chapter 4 and Chapter 6, respectively) were dominant on the egg surfaces
fertilized and reared in 0.2 µm filtered water.
Our study contributes to the broader literature on microbial community assembly and
succession by demonstrating that both local deterministic processes and dispersal play roles in
shaping the microbial communities assembly on the egg surfaces. PCA separation of egg surface
microbial communities from water microbial communities indicates that deterministic processes
on the egg surfaces are more dominant than dispersal. These dominant processes occurred fairly
quickly within 6 hours. We believe that these processes can be mediated via lysozyme and/or
abilities of microbes to adhere to eggs. It is important for us to study sturgeon egg chemistry in
the future to further elucidate the mechanisms shaping this early stage of host-microbe
interaction.

151

References

152

References

1.

Preston FW (1960) Time and Space and the Variation of Species. Ecology 41: 611-627

2.

Hall N (2007) Advanced sequencing technologies and their wider impact in microbiology.
Journal of Experimental Biology 210: 1518-1525

3.

Zengler K, Toledo G, Rappé M, Elkins J, Mathur EJ, Short JM, Keller M (2002)
Cultivating the uncultured. Proceedings of the National Academy of Sciences of the
United States of America 99: 15681-15686

4.

Huber JA, Mark Welch DB, Morrison HG, Huse SM, Neal PR, Butterfield DA, Sogin
ML (2007) Microbial Population Structures in the Deep Marine Biosphere. Science 318:
97-100

5.

Pommier T, Canback B, Riemann L, Bostrom KH, Simu K, Lundberg P, Tunlid A,
Hagstrom Å (2007) Global patterns of diversity and community structure in marine
bacterioplankton. Molecular Ecology 16: 867-880

6.

Van der Gucht K, Vandekerckhove T, Vloemans N, Cousin S, Muylaert K, Sabbe K,
Gillis M, Declerk S, De Meester L, Vyverman W (2005) Characterization of bacterial
communities in four freshwater lakes differing in nutrient load and food web structure.
FEMS Microbiology Ecology 53: 205-220

7.

Yannarell AC, Triplett EW (2005) Geographic and environmental sources of variation in
lake bacterial community composition. Applied and Environmental Microbiology 71:
227-239

8.

Fierer N, Jackson RB (2006) The diversity and biogeography of soil bacterial
communities. Proceedings of the National Academy of Sciences of the United States of
America 103: 626-631

9.

Garbeva P, Postma J, Van Veen JA, Van Elsas JD (2006) Effect of above-ground plant
species on soil microbial community structure and its impact on suppression of
Rhizoctonia solani AG3. Environmental Microbiology 8: 233-246

10.

Redford A, Fierer N (2009) Bacterial Succession on the Leaf Surface: A Novel System
for Studying Successional Dynamics. Microbial Ecology 58: 189-198

11.

Lambais MR, Crowley DE, Cury JC, Bull RC, Rodrigues RR (2006) Bacterial diversity
in tree canopies of the Atlantic forest. Science 312: 1917-1917

12.

Ley RE, Lozupone CA, Hamady M, Knight R, Gordon JI (2008) Worlds within worlds:
evolution of the vertebrate gut microbiota. Nature Reviews Microbiology 6: 776-788
153

13.

Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, Sargent M, Gill SR,
Nelson KE, Relman DA (2005) Diversity of the Human Intestinal Microbial Flora.
Science 308: 1635-1638

14.

Hutchinson GE (1957) Population studies - Animal ecology and demography Concluding Remarks. Cold Spring Harbor Symposia on Quantitative Biology 22: 415427

15.

Hubbell SP (2001) The unified neutral theory of biodiversity and biogeography.
Princeton University Press, Princeton, New Jersey, USA

16.

Leibold MA, McPeek MA (2006) COEXISTENCE OF THE NICHE AND NEUTRAL
PERSPECTIVES IN COMMUNITY ECOLOGY. Ecology 87: 1399-1410

17.

Dumbrell AJ, Nelson M, Helgason T, Dytham C, Fitter AH (2009) Relative roles of niche
and neutral processes in structuring a soil microbial community. ISME J 4: 337-345

18.

Van der Gucht K, Cottenie K, Muylaert K, Vloemans N, Cousin S, Declerck S, Jeppesen
E, Conde-Porcuna JM, Schwenk K, Zwart G (2007) The power of species sorting: local
factors drive bacterial community composition over a wide range of spatial scales.
Proceedings of the National Academy of Sciences of the United States of America 104:
20404-20409

19.

Palmer C, Bik EM, DiGiulio DB, Relman DA, Brown PO (2007) Development of the
human infant intestinal microbiota. PLoS Biology 5: e177

20.

Favier CF, Vaughan EE, De Vos WM, Akkermans ADL (2002) Molecular monitoring of
succession of bacterial communities in human neonates. Applied and Environmental
Microbiology 68: 219-226

21.

Fierer N, Hamady M, Lauber CL, Knight R (2008) The influence of sex, handedness, and
washing on the diversity of hand surface bacteria. Proceedings of the National Academy
of Sciences of the United States of America 105: 17994-17999

22.

Jackson CR, Churchill PF, Roden EE (2001) Successional changes in bacterial
assemblage structure during epilithic biofilm development. Ecology 82: 555-566

23.

Lyautey E, Jackson C, Cayrou J, Rols J-L, Garabétian F (2005) Bacterial Community
Succession in Natural River Biofilm Assemblages. Microbial Ecology 50: 589-601

24.

Anderson-Glenna MJ, Bakkestuen V, Clipson NJW (2008) Spatial and temporal
variability in epilithic biofilm bacterial communities along an upland river gradient.
FEMS Microbiology Ecology 64: 407-418

25.

Martiny AC, Jorgensen TM, Albrechtsen H-J, Arvin E, Molin S (2003) Long-term
succession of structure and diversity of a biofilm formed in a model drinking water
distribution system. Applied and Environmental Microbiology 69: 6899-6907

154

26.

Székely A, Sipos R, Berta B, Vajna B, Hajdú C, Márialigeti K (2009) DGGE and TRFLP analysis of bacterial succession during mushroom compost production and
sequence-aided T-RFLP profile of mature compost. Microb Ecol 57: 522-533

27.

Okabe S, Odagiri M, Ito T, Satoh H (2007) Succession of sulfur-oxidizing bacteria in the
microbial community on corroding concrete in sewer systems. Applied and
Environmental Microbiology 73: 971-980

28.

Fierer N, Nemergut D, Knight R, Craine JM (2010) Changes through time: integrating
microorganisms into the study of succession. Research in Microbiology 161: 635-642

29.

Finlay BJ, Clarke KJ (1999) Ubiquitous dispersal of microbial species. Nature 400: 828828

30.

Cherr GN, Clark WH (1982) Fine Structure of the Envelope and Micropyles in the Eggs
of the White Sturgeon, Acipenser transmontanus Richardson. Development, Growth &
Differentiation 24: 341-352

31.

Peterson D, Vecsei P, Jennings C (2007) Ecology and biology of the lake sturgeon: a
synthesis of current knowledge of a threatened North American Acipenseridae. Reviews
in Fish Biology and Fisheries 17: 59-76

32.

Lamy D, Obernosterer I, Laghdass M, Artigas LF, Breton E, Grattepanche JD, Lecuyer E,
Degros N, Lebaron P, Christaki U (2009) Temporal changes of major bacterial groups
and bacterial heterotrophic activity during a Phaeocystis globosa bloom in the eastern
English Channel. Aquatic Microbial Ecology 58: 95-107

33.

Sekiguchi H, Watanabe M, Nakahara T, Xu B, Uchiyama H (2002) Succession of
bacterial community structure along the changjiang river determined by Denaturing
Gradient Gel Electrophoresis and clone library analysis. Applied and Environmental
Microbiology 68: 5142-5150

34.

Kline KA, Fälker S, Dahlberg S, Normark S, Henriques-Normark B (2009) Bacterial
Adhesins in Host-Microbe Interactions. Cell host & microbe 5: 580-592

35.

Rendueles O, Ghigo J-M (2012) Multi-species biofilms: how to avoid unfriendly
neighbors. FEMS Microbiology Reviews 36: 972-989

36.

Kudo S (2000) Enzymes responsible for the bactericidal effect in extracts of vitelline and
fertilisation envelopes of rainbow trout eggs. Zygote 8: 257-265

37.

Saurabh S, Sahoo PK (2008) Lysozyme: an important defence molecule of fish innate
immune system. Aquaculture Research 39: 223-239

38.

Braun MH, Steele SL, Ekker M, Perry SF (2009) Nitrogen excretion in developing
zebrafish (Danio rerio): a role for Rh proteins and urea transporters. American Journal of
Physiology - Renal Physiology 296: F994-F1005

155

39.

Chadwick T, Wright P (1999) Nitrogen excretion and expression of urea cycle enzymes
in the atlantic cod (Gadus morhua l.): a comparison of early life stages with adults. The
Journal of Experimental Biology 202: 2653-2662

40.

Barton AD, Dutkiewicz S, Flierl G, Bragg J, Follows MJ (2010) Patterns of Diversity in
Marine Phytoplankton. Science 327: 1509-1511

41.

Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS,
McGarrell DM, Marsh T, Garrity GM, Tiedje JM (2009) The Ribosomal Database
Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Research
37: D141-D145

42.

Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naïve Bayesian Classifier for Rapid
Assignment of rRNA Sequences into the New Bacterial Taxonomy. Applied and
Environmental Microbiology 73: 5261-5267

43.

Liu W, Marsh T, Cheng H, Forney L (1997) Characterization of microbial diversity by
determining terminal restriction fragment length polymorphisms of genes encoding 16S
rRNA. Applied and Environmental Microbiology 63: 4516-4522

44.

Marsh TL (1999) Terminal restriction fragment length polymorphism (T-RFLP): An
emerging method for characterizing diversity among homologous populations of
amplification products. Current Opinion in Microbiology 2: 323-327

45.

R Development Core Team (2009) R: A launguage and environment for statistical
computing. R foundation for Statistical Computing, Vienna, Austria.

46.

Colombo RE, Garvey JE, Wills PS (2007) A guide to the embryonic development of the
shovelnose sturgeon (Scaphirhynchus platorynchus), reared at a constant temperature.
Journal of Applied Ichthyology 23: 402-410

47.

Kudo S, Teshima C (1991) Enzyme activities and antifungal action of fertilization
envelope extract from fish eggs. Journal of Experimental Zoology 259: 392-398

48.

Kudo S, Inoue M (1989) Bacterial action of fertilization envelope extract from eggs of
the fish Cyprinus carpio and Plecoglossus altivelis. Journal of Experimental Zoology 250:
219-228

49.

Fuda H, Hara A, Yamazaki F, Kobayashi K (1992) A peculiar immunoglobulin M (IgM)
identified in eggs of chum salmon (Oncorhynchus keta). Developmental and
Comparative Immunology 16: 415-423

50.

Van Veen W, Mulder E, Deinema MH (1978) The Sphaerotilus-Leptothrix group of
bacteria. Microbiological Reviews 42: 329

51.

Steele SL, Chadwick TD, Wright PA (2001) Ammonia detoxification and localization of
urea cycle enzyme activity in embryos of the rainbow trout (Oncorhynchus mykiss) in

156

relation to early tolerance to high environmental ammonia levels. J Exp Biol 204: 21452154
52.

Terjesen BF, Finn RN, Norberg B, Rønnestad I (2002) Kinetics and fates of ammonia,
urea, and uric acid during oocyte maturation and ontogeny of the Atlantic halibut
(Hippoglossus hippoglossus L.). Comparative Biochemistry and Physiology Part A:
Molecular and Integrative Physiology 131: 443-455

53.

Fraune S, Augustin R, Anton-Erxleben F, Wittlieb J, Gelhaus C, Klimovich VB,
Samoilovich MP, Bosch TCG (2010) In an early branching metazoan, bacterial
colonization of the embryo is controlled by maternal antimicrobial peptides. Proceedings
of the National Academy of Sciences of the United States of America 107: 18067-18072

54.

Fraune S, Augustin R, Bosch TCG (2011) Embryo protection in contemporary
immunology: Why bacteria matter. Communicative and Integrative Biology 4: 369-372

55.

Besemer K, Singer G, Limberger R, Chlup A-K, Hochedlinger G, Hödl I, Baranyi C,
Battin TJ (2007) Biophysical Controls on Community Succession in Stream Biofilms.
Applied and Environmental Microbiology 73: 4966-4974

157

CHAPTER 6: CHARACTERIZATION OF BACTERIAL ISOLATES FROM THE EGG
SURFACES OF LAKE STURGEON (ACIPENSER FULVESCENS) FOR
ANTAGONISTIC INTERACTIONS AND BIOFILM FORMING CAPABILITIES

Abstract
Microbes often interact in antagonistic ways, with some microbes displaying aggressive
tendencies over other susceptible microbes. Such interactions also relate to biofilm formation by
microbes, which is one mechanism that can be used to defend against antimicrobial activity.
Analyzing such interactions can help understand microbial community formation and may also
assist in identifying potential microbes that can be used as probiotic treatment to protect against
aggressors. In this study, we analyzed antagonistic interactions and biofilm forming capability of
25 representative isolates from the egg surface of the threatened Lake Sturgeon (Acipenser
fulvescens) using soft agar overlay and a crystal violet biofilm assay. Eight isolates exhibited
aggression to at least one other isolate. Antagonistic interactions were dependent on temperature
and phylogeny. Pseudomonas sp. C22 was found to be the most aggressive strain of all,
inhibiting growth of 15 out of the 25 isolates. Flavobacterium spp. were found to be one of the
least aggressive, and one of the most susceptible genera in the community. Four strong
aggressors were tested against 6 known fish pathogens. Each pathogen was inhibited by at least
one of the 4 aggressors. Pseudomonas sp. C22, which displayed the highest aggressiveness in
our study, inhibited growth of 5 out of the 6 fish pathogens. Strong biofilm forming capabilities
were observed for 11 isolates and they were dependent on environmental and nutritional

158

conditions. Hydrogenophaga sp. and Caulobacter sp., two of the most sensitive isolates to
antagonists, were among the best biofilm formers as was the strong aggressor Pseudomonas sp.
C22. Our results revealed the potential for a complex nature of interactions amongst members of
a microbial community on the eggs surface. This study was also significant in revealing
microbial populations with potential as a probiotic. The eight antimicrobial producing strains
isolated from the eggs surface may provide protection from pathogens.

159

Introduction
One of the central questions in microbial ecology addresses the nature of interactions
occurring among microbes in a complex community. These interactions span the range from
highly mutualistic to highly antagonistic. Antagonistic microbial interactions have been studied
in different microbial communities including communities associated with sponge, coral leaf,
and marine water [1-4]. These studies revealed phylogenetic trends in antagonistic interactions,
where some genera aggressively inhibited growth of others [1-3, 5]. These antagonistic
interactions were not static, but rather dynamic and influenced by environmental variables such
as temperature [4, 6], exposure to oxygen [7-8], biofilm formation [7], and nutrient level [8]. One
specific type of antagonistic interaction that has been studied extensively in the past involves
antagonistic interactions among host associated microbes against host associated pathogens [4, 6,
9-10]. Identifying such interactions has important implications for potential probiotic application
in systems in which control of pathogens is a priority [4, 6, 9-10].
There are a number of remaining unanswered questions in the realm of microbial
interactions. Although extensive research has been conducted to characterize individual isolates
for their biochemical and physiological properties, little is known regarding how properties of
individual isolates translate into broader functional or behavioral activities within complex
microbial communities that exist in nature. It is therefore important to characterize interactions
among microbes as members of broader communities and to approach the study of antimicrobial
activities from a community perspective [11-12]. Such an approach could yield new information
about fundamental ecological processes within a structurally complex community.
In a complex community frequently these interactions between populations play out
within the context of a biofilm. Microbes have been frequently observed to form biofilm on

160

surfaces of substrates in natural environments [12]. Relevant to interactions between populations
are observations that some antimicrobial substances are expressed only in biofilm [7], and a
biofilm formation allows some susceptible microbes to escape from antimicrobial activities [1314]. Biofilm is also formed to avoid host immune systems, which allows microbes to persist in
the host environment [15]. Thus, it is important to simultaneously analyze antagonistic
interactions and biofilm formation in order to fully understand interactions among microbes and
subsequent microbial community assembly.
The study of antimicrobial interactions and biofilm formation is important for the case of
the threatened fish species Lake Sturgeon (Acipenser fulvescens). Lake Sturgeon have
historically been a valuable asset to Michigan fishermen, most notably for their caviar. Declines
in the Lake Sturgeon populations due to over harvesting and habitat degradation [16-17] and low
natural recruitment [16] have garnered the attention of the scientific community. Previous studies
identified the high egg mortality of the sturgeon [18]. Since fish eggs extruded by female
sturgeon are rapidly colonized by a diverse collection of aquatic microbes, the role of microbial
communities in influencing egg mortality is of particular interest.
Our previous work identified significant associations between egg mortality and both
microbial quantity and community structure surrounding the egg surface (Chapter 3). We also
found that local deterministic processes such as host innate immunity, metabolite secretion, and
microbe-microbe interactions around the egg micro-environment are key processes shaping the
egg surface microbial communities (Chapter 4, 5). This current study builds on previous research
by investigating the role of microbe-microbe interactions that may potentially alter the microbial
community, and in turn affect mortality of the fish embryo.

161

In this study, we tested antagonistic interactions and biofilm forming capability among 25
representative isolates from the egg surface of the Lake Sturgeon at two temperature regimes,
which are relevant to the temperature range they experience during spawning season. We
hypothesized that isolates competing for resources and space in the microbial community may
exert some antimicrobial activity and demonstrate biofilm forming abilities in order to promote
their own persistence. This study is significant for the management of this threatened species in
particular because it will identify potential symbionts for probiotic treatment. The study is also
significant in contributing to a broader understanding of microbial community formation in
elucidating potential mechanisms of microbial community turnover during embryogenesis.

Methods
Study site and sample collection
Microbes were isolated from the surfaces of Lake Sturgeon eggs from the Black Lake
Population in Onaway, Michigan. Two types of fertilized eggs were included in the study; those
collected directly from spawning stream and those fertilized and reared in a streamside hatchery.
th

th

Stream eggs were collected on May 7 and May 17 in 2009 from two different sites in the
Upper Black River, which is the sole spawning stream for the Black Lake Population of Lake
Sturgeon. Eggs were collected from the bottom of the stream in a net with bottom substrates and
were subsequently removed using sterile tweezers. Eggs were rinsed with phosphate buffered
saline (PBS) and each egg was placed in a 2mL eppendorf tube with sterile PBS. The eppendorf
tubes were placed in a cooler box and were transported to the streamside hatchery within 30
minutes.

162

The hatchery-fertilized eggs were collected in 2010. Streamside hatchery is located at the
riverside of the Upper Black River in Onaway, MI. The hatchery water was pumped up from the
spawning stream. Large particulate matter in the stream water was removed using sock filters
and the filtered stream water was gravity fed to the hatchery system. Gametes collected from
spawning adults were fertilized and reared under different conditions in the hatchery for
production. The hatchery eggs were collected from fertilized eggs of three different families
reared in two water types (UV radiation-treated and untreated) and at two temperature regimes
o

o

(12 C and 19 C). The collected eggs were rinsed with PBS and placed in a 2mL eppendorf tube
with PBS.

Isolation of bacteria
Both stream-collected and hatchery-collected eggs were processed in the streamside
hatchery using the following procedures. Eggs in PBS were vortexed for 3 minutes and the
supernatant was diluted with PBS at different dilution factors. 100 µL of the diluted supernatant
were plated on R2A (0.5 g proteose peptone, 0.5 g casamino acids, 0.5 g yeast extract, 0.5 g
dextrose, 0.5 g soluble starch, 0.3 g dipotassium phosphate, 0.3 g sodium pyruvate, 0.05 g
Magnesium sulfate, and 15mg agar in 1 L MiliQ Water). The R2A plates were then incubated at
o

two different temperatures; one at 5 C in a refrigerator and the other at ambient temperature in
o

the hatchery (10 – 18 C). The plates with bacterial isolates were transported to a laboratory in
Michigan State University and then processed further. Each of the isolated colonies on R2A was
re-streaked on a fresh R2A plate, and an isolated colony on the plate was grown overnight in
o

o

R2B at room temperature in the laboratory (20 – 22 C) and stored at -80 C with a glycerol

163

concentration of 15%. Throughout the study, stream-collected isolates are identified by ID
numbers preceded by the letters A, B, C, D, or E and hatchery-fertilized isolates are identified by
ID numbers preceded by the letter F.

16S rRNA gene sequencing and phylogenetic analysis
Isolated strains were grown in 10 mL of R2B overnight and harvested by centrifugation
at 10,000 RPM in an SS34 rotor. Genomic DNA was extracted with a MoBio™ Power Soil
DNA extraction kit. The 16S rRNA gene was PCR amplified using “universal” primers 27F (5’AGA GTT TGA TCM TGG CTC AG - 3’) and 1389R (5’-ACG GGC GGT GTG TAC AAG 3’) and the resulting amplicons were purified with Qiagen™ PCR Cleanup columns. Purified
PCR products were sequenced at the Michigan State University Research Technology Support
Facility using an ABI 3730 capillary electrophoresis system with a 27F primer. The phylogenetic
relationships among the 92 egg isolates were inferred using MEGA version 4.0 [19] after
previously being aligned using Ribosomal Database Project (RDP) [20]. Phylogenetic
relationships were inferred using the Neighbor-Joining algorithm [21] with the Maximum
Composite Likelihood method [22]. A total of 471 informative sites (with complete deletion
option) within 16S rRNA gene were used to construct the Neighbor-Joining tree. At each node,
we calculated the frequency that the isolates appeared in the same cluster after conducting a 1000
bootstrap test.

Soft agar overlay technique for screening antagonists
We examined antagonistic interactions between 25 sturgeon egg isolates (Table 6.1).
These isolates were chosen from the total collection to best represent the phylogenetic diversity
of the egg surface microbial community. The 25 isolates were revived in 10 mL of R2B
164

o

overnight at room temperature (20 - 22 C) for 2 to 3 days. Isolates which could not be grown in
the broth were grown in R2A media at room temperature for 48 hours.

Table 6.1. List of isolates used for antagonistic interactions.
Assignment
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

Phylum/Class
Bacteroidetes
Bacteroidetes
Bacteroidetes
Bacteroidetes
Bacteroidetes
Bacteroidetes
Firmicutes
Gammaproteobacteria
Gammaproteobacteria
Gammaproteobacteria
Gammaproteobacteria
Gammaproteobacteria
Gammaproteobacteria
Gammaproteobacteria
Betaproteobacteria
Betaproteobacteria
Betaproteobacteria
Betaproteobacteria
Betaproteobacteria
Betaproteobacteria
Betaproteobacteria
Betaproteobacteria
Betaproteobacteria
Alpha proteobacteria
Deinococcus-Thermus

Genus/Species
Flavobacterium sp.
Flavobacterium sp.
Flavobacterium sp.
Flavobacterium sp.
Flavobacterium sp.
Flavobacterium sp.
Bacillus sp.
Aeromonas salmonicida
Aeromonas sobria
Aeromonas enchereia
Pseudomonas sp.
Pseudomonas sp.
Serratia sp.
Rheinheimera sp.
Massilia sp.
Janthinobacterium sp.
Janthinobacterium sp.
Iodobacter sp.
Deefgea sp.
Vogesella sp.
Undibacterium sp.
Hydrogenophaga sp.
Acidovorax sp.
Caulobacter sp.
Deinococcus sp.

165

Strain ID
A20
D11
E17
B30
C6
B10
C20
A12
A25
B25
D2
C22
D14
F1
B13
F8
F13
D4
D10
F3
F9
F14
F19
F16
F4

o

o

Antagonistic interactions among 25 isolates were screened at both 14 C and 20 – 22 C
by stamping 23 of the isolates on a soft agar lawn in which one of the 25 isolates was inoculated.
Isolates #1 and #5 were removed from the stamp since they exhibited gliding on the soft agar.
The soft agar overlay technique was previously described by Mendoza et al 1997 [23]. The
detailed procedure is as follows. The preparation of lawn for soft agar overlay was accomplished
by pouring 6 mL of soft R2A (one half of the agar concentration relative to the regular R2A
media) inoculated at 55 ºC with 100 µL of each broth culture in the equivalent phase of growth
for each trial (2-3 days of growth) or with 300 µL of plate culture suspension (3 loopfulls of
plate culture suspended in 1 mL of R2B) on top of a 20 mL pre-solidified regular R2A plate.
After cooling, the 23 non-gliding stamp strains were stamped in duplicate on the lawn using a 48
spike stamp and one half of a 96 well-plate filled with 200 µL of broth culture of each stamp
strain in each well. Three control plates with un-inoculated soft agar (beginning, middle, end)
were included for each trial to ensure that the stamp broth cultures were viable and stamped
correctly. The stamped plates including the three controls were then incubated at 20 – 22 ºC or
14 ºC for 2 days.
Antagonistic interactions were determined by the presence of zones of inhibition (zone of
clearance in soft agar) observed two days after stamping (Figure 6.1). The screening assay was
performed thrice with duplicates on each plate for a total of 6 replicates. Stamp strains which
exhibited antagonistic interactions with lawn strains in one of the 6 replicates were further
examined with a 12 replicate assay on a single plate to confirm the antagonistic interaction on a
one-on-one basis and to determine the extent of inhibition (Figure 6.1). The extent of antagonism
was quantified by measuring the radial zone of inhibition from the center of the stamped colony
to the point where the lawn strain’s normal opacity returned. Width of the zone of inhibition was
166

calculated subtracting the radius of colonies from the radial zone of inhibition. To investigate the
reciprocal relationship of antagonistic interactions, the size of colony of all 23 stamp strains were
o

o

measured for each lawn strain at both 14 C and 20 – 22 C after 4 days of incubation (the timing
o

o

of measurements were synchronized for both 14 C and 20 – 22 C).

Figure 6.1. Antagonistic interaction screening and confirmation using soft agar overlay
assay.
Left: Screening for antagonistic interactions using stamps. Right: Confirmation for antagonistic
interaction on a one-on-one basis with 12 replicates.

This procedure was repeated on a select group of the top 4 most aggressive of our 25
isolates with one of 6 known fish pathogens as the lawn strain. Fish pathogens included 3
Aeromonas spp. strains and 2 Flavobacterium spp. strains isolated from either inland lakes in
Michigan or the Great Lakes and Yersenia ruckeri ATCC 29473, another known fish pathogen.

167

Fish pathogens were provided by Dr. Mohamed Faisal and Dr. Tom Loch of Michigan State
University.

Biofilm assays
Biofilm forming capability of 23 isolates (excluding isolates #3 and #4 which were not
culturable in broth) was investigated through an adaptation of a 96-well plate assay with crystal
violet [24]. For each trial, 200 µL of culture broth of each of the 23 isolates were placed in
duplicate into one half of a 96-well plate. In addition to 23 isolates, Pseudomonas aeruginosa
was used as a positive control and a single blank well was used as a negative control. A 48-spike
stamp was used to transfer the isolate broth cultures and controls into wells filled with 175 µL of
sterile media in triplicate (a total of 6 replicates per isolate). Three nutrient media were tested in
the biofim assay: R2B (same as R2A above but without agar): M9+Glucose (12.8g Na2HPO4, 3g
KH2PO4, 0.5g NaCl, 1g NH4Cl, 2ml of 1M MgSO4, and 4g Glucose per liter); and
M9+Casamino Acids (M9 + 4g Casamino Acids per liter). The inoculated plates were incubated
at 14ºC or 20 – 22 ºC shaking at 100 rpm for 2 days. Post-incubation optical density at 600 nm
was measured to quantify growth within the broth. Cells adhering to the surfaces of the plate
were stained with 200 µL of 0.1% crystal violet dye for 15 minutes, rinsed in water baths and
then inverted to dry completely. The crystal violet stain was then extracted from each well with
the addition of 200 µL of 30% acetic acid for 15 minutes. The entire dye solution within each
well was transferred into new plates and the absorbance of the dye was measured at 600 nm
using a spectrophotometer.

168

Results
Isolation and phylogenetic affiliation
We collected 92 isolates from the egg surface of Lake Sturgeon (Figure 6.2). Our culture
collection covered a range of the egg surface microbial community diversity that was previously
characterized using 16S rRNA gene pyrosequencing. Our culture collection contained a broad
spectrum of bacteroidetes and beta proteobacteria on the egg surface, but was underrepresented
in the -proteobacteria. 25 isolates out of the 92 that represent the egg surface microbial
community assembly were chosen and tested for both antagonistic interactions and biofilm
forming capabilities.

Antagonistic interactions between isolates
Among the 25 strains we tested, 8 isolates (30%) were confirmed to be positive for antagonisms
to at least one of the other strains at one of the tested temperatures (Table 6.2). The incidence of
o

antagonistic interaction was dependent on temperature. At 21 C, isolates that exhibited
aggressiveness to at least one other isolate (aggressor) increased their target range (Figure 6.3)
o

and intensity of antagonism relative to results from 14 C (Table 6.2). The number of aggressors
increased by one from the low to high temperature regime, as Bacillus sp. showed aggressiveness
o

only at 21 C. Susceptible isolates increased from 12 to 18 as the incubation temperature
increased (Table 6.2, Figure 6.3). Six isolates including Flavobacterium sp. B10 and
o

Caulobacter sp. F16 exhibited susceptibility only at 21 C. Overall, 50 to 70 % (dependent on
temperature) of the isolates in the community were sensitive to at least one of the isolates.

169

Figure 6.2. Phylogenetic relationships among 92 sturgeon egg isolates inferred using the
Neighbor-Joining method.
25 isolates that were used in this study are highlighted red. The bootstrap values above 75 are
shown at nodes.

170

Table 6.2. Antagonistic interactions among 25 egg isolates at two different temperatures

Color code indicates the extent of inhibition as a unit of zone of inhibition from the edge of
colonies (mm). The bottom row in the table shows cumulative stamp success of the stamp strains
(aggressiveness).The right-most column summarizes the sensitivity of lawn strains to stamp
strains (sensitivity). Flavobacterium sp. A20 and C6 were removed from the stamp due to
gliding ability.

171

Table 6.2. (cont’d).

172

The antagonistic interactions displayed a phylogenetic trend (Table 6.2). Genus
Flavobacterium was found to be the least aggressive genus and one of the most susceptible
genera in the community. The majority of -proteobacteria were also susceptible to antagonistic
interactions. Bacillus sp. and -proteobacteria including genus Pseudomonas inhibited growth of
Flavobacterium isolates and isolates of -proteobacteria. Seven out of nine -proteobacteria
o

were susceptible to one of the -proteobacteria isolates at 21 C and 5 out of six Flavobacterium
isolates were susceptible to both Bacillus sp. and one of the -proteobacteria isolates. Genus
Pseudomonas and Bacillus were found to be strongly aggressive. At a strain level analysis, the
most aggressive isolate was found to be Pseudomonas sp. C22 which inhibited growth of 15
isolates out of 25. The most susceptible strain was Hydrogenophaga sp. F14 which was
susceptible to 7 isolates, some with a large zone of inhibition (Table 6.2).
However, Some isolates deviated from the phylogenetic trend. Although about 80% of
beta-proteobacteria isolates were sensitive to at least one of the other isolates, strains Vogesella
sp. F3 and Janthinobacterium sp. F8 showed aggressiveness to other isolates and were not
susceptible to any aggressors. The strain Flavobacterium sp. B10 was also resistant to all
aggressors.

173

Figure 6.3. Summary of (a) lawn susceptibilities and (b) stamp aggressiveness among 25
sturgeon egg surface isolates at two temperature regimes
174

The antagonistic interaction held in the reciprocal condition when positions were
switched from stamp to lawn. Evidence of aggression was detected when the aggressor was the
lawn strain in the soft agar in that the size of the sensitive stamped-strains were diminished
(Figure 6.4). This negative effect of lawn isolates on growth of stamp isolates was detected even
with non-antagonistic pairs of isolates, suggesting the existence of resource competition among
the isolates (Figure 6.5).
The four strong aggressors Bacillus sp. C20, Pseudomonas sp. C22, Serratia sp. D14,
Janthinobacterium sp. F8 were tested against six known fish pathogens. All four aggressors
inhibited at least one of the 6 fish pathogens, and all of the 6 fish pathogens were inhibited by at
least one of the four aggressors (Figure 6.6). Pseudomonas sp. C22, which displayed the highest
aggressiveness in our study, inhibited growth of 5 out of the 6 fish pathogens. The most
susceptible fish pathogen was Flavobacterium sp. C05, whose growth was inhibited by 3 out of
the 4 aggressors we tested. In contrast to Flavobacterium sp. C05, F. columnare 090702-1 was
susceptible to only one aggressor we tested.

175

Figure 6.4. Reciprocal antagonistic interactions among 25 egg surface isolates when
positions are switched from stamp to lawn.
Isolates that acted as strong aggressors when used as stamps inhibited growth of stamp colonies
when used as lawn isolates. The figure depicts the negative correlation between aggressiveness
and the average growth inhibition.

176

Figure 6.5. Reduction of colony size due to both active growth inhibition by lawn isolate and resource competition under
two temperature regimes.

177

Figure 6.5. (cont’d).

178

Figure 6.6. Antagonistic interactions against 6 known fish pathogens by the 4 most
aggressive isolates identified on the sturgeon egg surface.

Biofilm forming capability
Using the crystal violet assay for biofilm formation, strong biofilm forming capabilities
(with average absorbance over 0.5) were observed for 11 isolates (nearly half of the total) and
they were both temperature and nutrient medium dependent (Figure 6.7). The pattern of biofilm
formation and temperature/medium dependency was strain dependent. Flavobacterium sp. A20
o

and D11 formed biofilm at 21 C when grown on minimum medium plus glucose as did three 179

proteobacteria isolates (Pseudomonas sp. D2, Pseudomonas sp. C22, and Serratia sp. D14). Two
-proteobacteria isolates (Hydrogenophaga sp. F14 and Acidovorax sp. F19) formed biofilm
o

when grown on either R2B or minimum medium plus casamino acid at 21 C.

There was also a relationship between planktonic growth in broth medium and biofilm
formation (Figure 6.8). However, planktonic growth did not always directly correlate with
biofilm forming capability. For instance, Flavobacterium sp. A20 and D11 grew well in both
R2B and minimum medium plus glucose, but biofilm formation was detected only in the
minimum medium plus glucose. In addition, planktonic growth in broth medium was not a
necessary condition for biofilm formation. For instance, Pseudomonas sp. D2 did not grow
o

planktonically at 14 C in minimum medium plus glucose but displayed the maximum biofilm
formation under this same condition.
There was a relationship between biofilm forming capacities and antagonistic interactions.
Hydrogenophaga sp. F14 and Caulobacteter sp. F16 were sensitive to antagonistic interactions
o

o

in R2A soft agar at 21 C and they were a good biofilm former in R2B at 21 C. On the other
hand, Pseudomonas sp. C22, the most aggressive isolate, was also found to be a good biofilm
former.

180

Figure 6.7. Quantification of biofilm formation with crystal violet in a microtiter plate assay under different media
181

Figure 6.8. Quantification of planktonic growth in different nutrient broth with OD600 in a microtitier plate assay.
182

Discussion
This study identified antagonistic interactions among isolates from the egg surface of the
Lake Sturgeon. The significant effect of temperature on antagonistic interactions reveals the
complex nature of microbial interactions on the egg surface microbial community. Temperature
dependency of antimicrobial interactions has been previously reported in other systems [4, 6],
but has particular relevance to the sturgeon host due to the temperature-dependent spawning
behavior in this species. Adult sturgeon spawn at two distinct time points during the spawning
season, which normally correspond to stream temperature. Our results suggest that it is possible
that different microbial interactions in the communities may be occurring at the different time
periods as a result of temperature, which could potentially result in differences in the egg
microbial community structure and egg mortality.
The phylogenetic trends we observed in antagonistic interactions were revealing in
highlighting the evolutionary origins of microbial interactions. Previous studies by others found
a similar phylogenetic trend in antagonistic interactions in other microbial communities [1-3, 5].
Studies on microbial community in marine water found that phylum Bacteroidetes was the least
aggressive and the most susceptible to antagonistic interactions [5], and Flavobacterium spp.
were the least aggressive genus in the community [1]. Strong inhibitory activities by Bacillus spp.
[1] and γ-proteobacteria including Pseudomonas spp. [3] were also reported in the marine
organic aggregates community and the sponge associated microbial community, respectively.
However, in our study not all strains within the same genus or class follow the phylogenetic
trend, suggesting that generalization in genus or class level do not always hold. Other studies
also found that some antagonistic interactions were strain specific [2].

183

Another important finding of this study dealt with the significant effect of resource
competition on microbe interactions, irrespective of antagonism. Resource competition is rarely
reported in studies on antagonistic interactions, but potentially affects the microbial community
structure and functional processes in the community. The negative effect of lawn isolates on
growth of all isolates suggests that future studies should also examine this phenomenon of
resource competition among members of a community given limited resources.
The significant antagonism displayed by the aggressor isolates against the known fish
pathogens suggests that there is potential for microbes to help protect fish eggs against invasion
of pathogenic bacteria. Pseudomonas sp. C22, was the most aggressive isolate of all of those we
tested, since it inhibited 5 out of the 6 known fish pathogens, and should be strongly considered
for potential future use in probiotic treatment for this threatened fish species. In fact,
Pseudomonas spp. have been used as probiotics to protect other fish species and have been found
to be effective [25-26].
Another important finding of this study concerned biofilm formation capabilities. Biofilm
formation is relevant to microbial interactions since susceptible isolates can persist in a
community by escaping from antimicrobial substances in the biofilm [13-14]. The dependence of
biofilm formation on both temperature and nutrient medium highlighted the complex nature of
biofilm formation, which corroborates findings in other systems [27-28]. Our findings together
with observations that egg metabolites can change during embryogenesis [29] suggest that
biofilm formation on the egg surfaces may be altered during embryogenesis. Such complexity
can also be appreciated through the finding that some isolates showed a small planktonic growth
but a large biofilm formation under the same condition, suggesting preference to form biofilm
rather than grow planktonically.

184

It is also interesting to note that we found some correlations between antagonistic
interactions and biofilm formation in some strains. For instance, Hydrogenophaga sp. F14 and
Caulobacteter sp. F16 were among the most susceptible strains and also were substantial biofilm
formers. They may persist in the egg surface microbial community by evading the antagonistic
interactions through biofilm formation as previously reported by other studies [13-14]. On the
other hand, we also found a biofilm capability of a strong aggressor, specifically Pseudomonas
sp. C22. This strain may use biofilm formation as a mechanism to allow persistence in the egg
surface microbial community. This finding again highlights the fact that this strain may be a
good candidate for probiotic treatment.
The antagonistic interactions we observed could also help explain the community
turnover we observed in our earlier studies through 454 pyrosequencing of egg surface microbial
community during embryogenesis (Chapter 2, 3, and 5). The pryosequencing analysis revealed
that the genera Acidovoras and Caulobacter persisted in the communities throughout
embryogenesis, genus Hydrogenophaga increased its relative abundance toward the late egg
developmental stage, and genus Flavobacterium decreased toward the late egg developmental
stage. Our biofilm assays demonstrated that genus Acidovorax, Caulobacter, and
Hydrogenophaga were capable of forming a biofilm, which might allow them to persist in the
microbial community on the egg surface. The fact that Flavobacterium was one of the most
susceptible genera to antagonistic interactions could contribute to the decline of this microbe in
the community toward the later period of egg development. Our previous work using
pyrosequencing also revealed that Pseudomonas spp. persisted in the community throughout
embryogenesis, but did not become dominant in the community. Although we expected that
Pseudomonas would increase in relative abundance in the community toward the end of

185

embryogenesis by utilizing antagonistic interactions and biofilm formation, there may be some
unknown mechanisms that inhibit the dominance of Pseudomonas in the community.
In this study, we did not investigate the underlying mechanisms behind the antagonistic
interactions, but previous studies by others offer insights into potential mechanisms. For instance,
mechanisms of active growth inhibition by Bacillus spp. [30-31] and Pseudomonas spp. [32-33]
have been studied including purification of antimicrobial substances and their induction
mechanisms. We also found that two strains in beta-proteobacteria Vogesella sp. F3 and
Janthinobacterium sp. F8 exhibited antimicrobial activities, however, the mechanisms of these
antagonistic interactions have not been previously characterized. Future research should be
directed toward understanding the mechanisms behind these antagonisms, specifically
purification of antibacterial substances.
We do not know how our findings translate into the natural setting where microbes
interact on the egg surface in a stream. Although we mimicked temperature ranges eggs
experienced during the spawning season, the nutrient content, concentrations, and physical
structure of the R2A plate are different from the egg surface environment. Our antagonistic
interactions represent those between a pure colony and a pure cell aggregate (lawn), which may
differ from the situation on the egg surfaces where the density of microbes is more scattered.
Furthermore, we tested the antagonistic interactions on a one-on-one basis, but in nature there
can be one against many or many against many relationships. We also incubated soft agar plates
in the atmosphere, while microbes in the natural setting are submerged in a stream. Previous
studies demonstrated that the level of oxygen affect the induction of antimicrobial substances [78]. Both nutrient uptake and delivery of antimicrobial substances could be affected by stream
flows. Although we tested biofilm forming capability using pure cultures in broth media, the

186

presence of antagonistic strains or antimicrobial substances can also induce biofilm formation
[34]. Future studies should seek to investigate these other avenues of research.
In summary, our results revealed the complex nature of microbial interactions amongst
the members of the egg surface microbial community. To our knowledge, this is the first study
on any system to simultaneously analyze antagonistic interactions and biofilm formation, an
approach which was informative in revealing significant patterns in microbial interactions.
Although our findings may not directly translate into microbial interactions observed on the egg
surfaces in the natural environment, our results suggest a potential use of Pseudomonas sp. C22
for probiotic application as a management tool to decrease egg mortality.

187

References

188

References

1.

Grossart H-P, Schlingloff A, Bernhard M, Simon M, Brinkhoff T (2004) Antagonistic
activity of bacteria isolated from organic aggregates of the German Wadden Sea. FEMS
Microbiology Ecology 47: 387-396

2.

Lo Giudice A, Brilli M, Bruni V, De Domenico M, Fani R, Michaud L (2007)
Bacterium–bacterium inhibitory interactions among psychrotrophic bacteria isolated from
Antarctic seawater (Terra Nova Bay, Ross Sea). FEMS Microbiology Ecology 60: 383396

3.

Mangano S, Michaud L, Caruso C, Brilli M, Bruni V, Fani R, Lo Giudice A (2009)
Antagonistic interactions between psychrotrophic cultivable bacteria isolated from
Antarctic sponges: a preliminary analysis. Research in Microbiology 160: 27-37

4.

Rypien KL, Ward JR, Azam F (2010) Antagonistic interactions among coral-associated
bacteria. Environmental Microbiology 12: 28-39

5.

Long RA, Azam F (2001) Antagonistic Interactions among Marine Pelagic Bacteria.
Applied and Environmental Microbiology 67: 4975-4983

6.

Long RA, Rowley DC, Zamora E, Liu J, Bartlett DH, Azam F (2005) Antagonistic
Interactions among Marine Bacteria Impede the Proliferation of Vibrio cholerae. Applied
and Environmental Microbiology 71: 8531-8536

7.

Yan L, Boyd KG, Adams DR, Burgess JG (2003) Biofilm-Specific Cross-Species
Induction of Antimicrobial Compounds in Bacilli. Applied and Environmental
Microbiology 69: 3719-3727

8.

Saha M, Ghosh D, Garai D, Jaisankar P, Sarkar KK, Dutta PK, Das S, Jha T, Mukherjee J
(2005) Studies on the production and purification of an antimicrobial compound and
taxonomy of the producer isolated from the marine environment of the Sundarbans.
Applied Microbiology and Biotechnology 66: 497-505

9.

Evans J, Armstrong T-N (2006) Antagonistic interactions between honey bee bacterial
symbionts and implications for disease. BMC Ecology 6: 4

10.

Marín-Cevada V, Muñoz-Rojas J, Caballero-Mellado J, Mascarúa-Esparza MA,
Castañeda-Lucio M, Carreño-López R, Estrada-de los Santos P, Fuentes-Ramírez LE
(2012) Antagonistic interactions among bacteria inhabiting pineapple. Applied Soil
Ecology 61: 230-235

11.

Hibbing ME, Fuqua C, Parsek MR, Peterson SB (2010) Bacterial competition: surviving
and thriving in the microbial jungle. Nature Reviews Microbiology 8: 15-25
189

12.

Rendueles O, Ghigo J-M (2012) Multi-species biofilms: how to avoid unfriendly
neighbors. FEMS Microbiology Reviews 36: 972-989

13.

Narisawa N, Haruta S, Arai H, Ishii M, Igarashi Y (2008) Coexistence of AntibioticProducing and Antibiotic-Sensitive Bacteria in Biofilms Is Mediated by Resistant
Bacteria. Applied and Environmental Microbiology 74: 3887-3894

14.

Burmølle M, Webb JS, Rao D, Hansen LH, Sørensen SJ, Kjelleberg S (2006) Enhanced
Biofilm Formation and Increased Resistance to Antimicrobial Agents and Bacterial
Invasion Are Caused by Synergistic Interactions in Multispecies Biofilms. Applied and
Environmental Microbiology 72: 3916-3923

15.

Anderson GG, Palermo JJ, Schilling JD, Roth R, Heuser J, Hultgren SJ (2003)
Intracellular Bacterial Biofilm-Like Pods in Urinary Tract Infections. Science 301: 105107

16.

Baker EA, Borgeson DJ (1999) Lake Sturgeon abundance and harvest in Black Lake,
Michigan, 1975-1999. North American Journal of Fisheries Management 19: 1080-1088

17.

Smith KM, Baker EA (2005) Characteristics of spawning lake sturgeon in the Upper
Black River, Michigan. North American Journal of Fisheries Management 25: 301-307

18.

Forsythe PS (2010) Exogenous correlates of migration, spawning, egg deposition and egg
mortality in the lake sturgeon (Acipenser fulvescens). Ph.D. Dissertation. Department of
Fisheries and Wildlife. Michigan State University. #3417681. pp191

19.

Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics
Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24: 15961599

20.

Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS,
McGarrell DM, Marsh T, Garrity GM, Tiedje JM (2009) The Ribosomal Database
Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Research
37: D141-D145

21.

Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing
phylogenetic trees. Molecular Biology and Evolution 4: 406-425

22.

Tamura K, Nei M, Kumar S (2004) Prospects for inferring very large phylogenies by
using the neighbor-joining method. Proceedings of the National Academy of Sciences of
the United States of America 101: 11030-11035

23.

Mendoza L, Wilkens M, Urzúa A (1997) Antimicrobial study of the resinous exudates
and of diterpenoids and flavonoids isolated from some Chilean Pseudognaphalium
(Asteraceae). Journal of Ethnopharmacology 58: 85-88

24.

Merritt JH, Kadouri DE, O'Toole GA (2011) Growing and Analyzing Static
BiofilmsCurrent Protocols in Microbiology. John Wiley & Sons, Inc.
190

25.

Goldschmidt-Clermont E, Wahli T, Frey J, Burr SE (2008) Identification of bacteria from
the normal flora of perch, Perca fluviatilis L., and evaluation of their inhibitory potential
towards Aeromonas species. Journal of Fish Diseases 31: 353-359

26.

Gobeli S, Goldschmidt-Clermont E, Frey J, Burr SE (2009) Pseudomonas chlororaphis
strain JF3835 reduces mortality of juvenile perch, Perca fluviatilis L., caused by
Aeromonas sobria. Journal of Fish Diseases 32: 597-602

27.

Rao T (2010) Comparative effect of temperature on biofilm formation in natural and
modified marine environment. Aquatic Ecology 44: 463-478

28.

Molobela I, Ilunga F (2012) Impact of bacterial biofilms: the importance of quantitative
biofilm studies. Annals of Microbiology 62: 461-467

29.

Chadwick T, Wright P (1999) Nitrogen excretion and expression of urea cycle enzymes
in the atlantic cod (Gadus morhua l.): a comparison of early life stages with adults. The
Journal of Experimental Biology 202: 2653-2662

30.

Mannanov RN, Sattarova RK (2001) Antibiotics Produced by Bacillus Bacteria.
Chemistry of Natural Compounds 37: 117-123

31.

Milner JL, Silo-Suh L, Lee JC, He H, Clardy J, Handelsman J (1996) Production of
kanosamine by Bacillus cereus UW85. Applied and Environmental Microbiology 62:
3061-3065

32.

James DW, Gutterson NI (1986) Multiple antibiotics produced by Pseudomonas
fluorescens HV37a and their differential regulation by glucose. Applied and
Environmental Microbiology 52: 1183-1189

33.

Haas D, Keel C (2003) Regulation of antibiotic production in root-colonizing
Pseudomonas spp. and relevance for biological control of plant disease. Annual Review
of Phytopathology 41: 117-153

34.

Hoffman LR, D'Argenio DA, MacCoss MJ, Zhang Z, Jones RA, Miller SI (2005)
Aminoglycoside antibiotics induce bacterial biofilm formation. Nature 436: 1171-1175

191

CHAPTER 7: CONCLUDING REMARKS

Factors affecting microbial community assembly on Lake Sturgeon eggs
Throughout this study, we demonstrated that microbial community assemblages on the
egg surfaces of the Lake Sturgeon were affected by a number of different factors. The most
important factor was found to be the local egg related factors, probably mediated by maternally
provisioned innate immunity, including lysozyme or egg surface chemistry, the latter of which
affects adhesion of microbes. This process separated the egg associated microbial communities
from the source water microbial communities within 6 hours of fertilization or even sooner
(Chapter 5, Chapter 6). The reason we believe that this process is mediated by lysozyme is that
phyla Actinobacteria and Firmicutes, both gram positive bacteria, were selected against by the
eggs (Chapter 2, Chapter 5, and Chapter 6). One of the lines of evidence that supports the
alternative adhesion theory is derived from the experiment in which we inoculated Acidovorax
and Flavobacterium during fertilization. Acidovorax successfully colonized the egg surfaces but
Flavobacterium did not, and the process happened within one hour. Although we cannot exclude
the possibility that this low Flavobacterium colonization was also mediated by lysozyme
(Flavobacterium which is gram negative is somehow sensitive to lysozyme), this initial adhesion
process is worthy of study in the future.
Our study revealed that another important process explaining microbial community
assembly was dispersal of aquatic microbes. Differences in the aquatic microbial community
affected the egg-associated microbial community. This effect was more significant in explaining
microbial community assembly than that of water temperature (Chapter 3). This process was also
dependent on the quantity of aquatic microbes. When the quantity of the aquatic microbial
192

community was lowered, the effect of aquatic microbial community on the egg surface microbial
community was diminished (Chapter 6). This density dependency is thus a fundamental aspect of
the nature of the effect of dispersal on microbial community assembly.
Furthermore, directional changes in the egg surface microbial community were observed
along with egg developmental stage. These changes were consistent throughout different
experiments and treatments (Chapter 2, Chapter 3, Chapter 5). These temporal changes in the
microbial community could be mediated by a number of different factors, including changes in
metabolites secreted by eggs, changes in lysozyme from types that are maternally provisioned to
those secreted by eggs, antagonistic microbial interactions, biofilm formation (Chapter 6) and
growth. We do not know the relative contributions of each factor, but this avenue of research is
worthy of pursuing in the future.
Temperature was another important factor explaining the egg surface microbial
community. The effect of temperature on microbial community assembly is not entirely
surprising, since it has been shown that microbes have preferences for certain temperature ranges,
with some microbial species growing better in warm temperature and others growing better in
cold. There was also an interaction between temperature and antagonistic interactions (Chapter
6). This suggests that antagonistic interactions act differently in their effects on microbial
communities at different temperatures. However, while temperature did affect the egg microbial
community, the effect was not as significant as that of the aquatic microbial community and the
temporal effect in the PCA analysis (Chapter 3). In addition, we did not detect an effect of
temperature on microbial quantity or egg mortality in the temperature ranges we studied.
However, water temperature did affect the incubation period of embryos and the larvae size at
hatch (Chapter 3). In summary, temperature is an important factor to consider when rearing

193

sturgeon eggs, since it affects not only the egg surface microbial composition, but also the life
history traits of the host.
Flow rate was also a factor that significantly affected microbial community composition.
High flow rate, as simulated in the flume experiment, lowered the diversity of the microbial
community on the egg surfaces, and the effect of flow on the egg associated microbial
community was observed only in the middle stage of embryogenesis (Chapter 2). We also found
that the lower the flow, the lower the abundance of the genus Flavobacterium on the egg surface
(Chapter 2). We did not pursue the effect of flow rate on host life history traits including egg
mortality and larvae size at hatch, so this effect should be further investigated in the future.
We also studied antagonistic interactions and biofilm forming capabilities among
sturgeon egg isolates. We found that both antagonistic interactions and biofilm forming
capability of sturgeon egg isolates were complex and dependent on temperature (and nutrient
composition for the case of biofilm formation, Chapter 6). However, we do not know how much
the microbial interactions or biofilm formation contribute to the observed changes in the egg
microbial community structure during incubation, a topic which should be further studied in the
future.
Our study also demonstrated that history may matter in explaining the future
development of a microbial community. Several lines of evidence suggest that the effect of
initial colonization on development of subsequent microbial communities was likely dependent
on both available space on eggs and/or the concentration of the source water microbes. When
eggs were inoculated with different strains for 60 minutes and were transferred and reared in the
stream water with high microbial load, the initial microbial community structure was masked and
converged with the community of stream fertilized and reared eggs (Chapter 4). However, when

194

the eggs fertilized in stream water (with high microbial load) for 6 hours were transferred and
reared in 0.2 µm filtered water, the egg associated microbial community did not converge with
that fertilized and reared in 0.2 µm filtered water. This suggested that either available space on
the egg surface or the concentration of the source water microbes plays a role in determining the
subsequent microbial community assembly. This suggests that the history of community
composition may be important particularly when a large impact (microbial load) occurred in the
initial stage and smaller impacts (microbial load) occurred in the later stages.

Factors affecting egg surface microbial quantity
5.5

Egg surface microbial quantity increased to the range of 10

6

to 10 16S rRNA gene

copies per egg within 6 hours post-fertilization (Chapter 5). Source water microbial quantity is
likely a factor mediating this process, since eggs fertilized in stream water had the highest
microbial quantity among those studied in the three water treatment. However, the effect of
source water on egg surface microbial quantity is also likely mediated by microbial composition.
When source water microbial composition was effectively controlled by eggs, as seen by the
lower abundance of phyla such as Actinobacteria, Firmicutes, and Bacteroidetes on egg surfaces,
less microbes will likely colonize the egg surface. In contrast, when composition is dominated by
microbes that are accepted by host eggs, including Acidovorax, the egg microbial community
likely increases even if the source water quantity is low.
We also detected potential growth of microbes on the egg surface. This was detected
when eggs were fertilized in stream water and transferred to and reared in 0.2 µm filtered water
(Chapter 5). The microbial quantity on eggs fertilized in stream water and reared in 0.2 µm
filtered water increased relative to the quantity of eggs fertilized and reared in 0.2 µm filtered
195

water. However, microbial community structure of these eggs (fertilized in stream and reared in
0.2 µm filtered water) did not converge with those fertilized and reared in 0.2 µm, which
suggests no effect of dispersal in explaining community structure. We believe that this increase
in microbial quantity can be attributed to the growth of microbes nested within the initial
colonizers.

The effect of microbes on the host life history traits
Our results regarding effects of microbial successional processes on host life history were
also illustrative in contributing to broadly understanding microbe-host interactions. Eggs
fertilized and reared in stream water (with high microbial loads) had significantly higher egg
mortality and smaller yolk sac areas than those reared in treated water (with lower microbial
loads) (Chapter 5). This suggests that yolk resources were being used against microbes that
colonized the egg surface. We also demonstrated that eggs fertilized with a putative symbiont
had lower egg mortality and less yolk resource use (Chapter 4). The fact that microbes affect
host life history traits such as larval size at hatch has significant implications by revealing the
extent of complexity in reciprocal interactions between microbes and hosts.

Symbionts and probiotic treatment
We also conducted preliminary work on identifying putative symbiotic microbial species
for the Lake Sturgeon eggs for potential future use in probiotic treatment to improve observed
high egg mortalities in this species. By rearing eggs in the presence of different compositions of
aquatic microbial communities and analyzing the correlation between presence of certain
microbes and the egg mortality, we identified a key set of microbial species that significantly

196

improved egg survival (Chapter 5). We also inoculated a putative symbiont to eggs during
fertilization, a treatment which reduced both egg mortality and resources used by treated
embryos (Chapter 4). This line of evidence suggests that this putative symbiont indeed improved
the fitness of the host, which is the broad definition of a symbiont. We also tested 25 egg isolates
representing the egg surface microbial community assembly for antagonistic interactions against
known fish pathogens (Chapter 6). Through this experiment, we identified several isolates that
are potentially beneficial to eggs by protecting them from known fish pathogens.

Egg microbiome
Prior to starting this project, we were wondering whether an egg microbiome exists.
Unlike the animal gut where space is contained and association with microbes can be long-term
6

in nature, eggs are exposed to a large number of aquatic microbes (an average of 10 cfu/ mL)
throughout incubation, which could be a random collection of microbes and prevent a distinct
microbiome from forming. However, our study suggests that it may still be possible for an egg
microbiome to form, since we demonstrated that egg microbial communities were distinct from
the source water microbial communities. Although we treated eggs with different water types, all
microbial communities associated with eggs were clustered together relative to the source water
(Chapter 5). It took approximately 6 hours post-fertilization for eggs to shape the communities
and form a microbiome, although eggs from one of the families shaped the egg associated
microbial community within 1 hour.
However, it is also important to note that there was some variation in the microbiome and
a number of factors that affected its formation. In the water experiment, microbial communities
were grouped by water type within the egg microbial community cluster (suggesting the effect of
197

dispersal on the egg microbiome) (Chapter 3 and 5). We also observed directional changes of
microbial communities along with the egg developmental stages across a number of experiments
(Chapter 2, Chapter 3, and Chapter 5). Temperature and flow rate also affected the microbial
communities on the egg surfaces. These findings suggest that egg microbiomes can vary with
water type, environmental variables, and egg developmental stages, although they are clustered
closely and significantly different from the source water community.

Future work
The findings of this dissertation can be used to develop hypotheses and formulate new
research questions for those who may further pursue this topic in the future.
One of the major areas we have not yet explored deals with the chemistry of the sturgeon
eggs. One of the largest gaps of information concerns the egg surface chemistry, which is a key
to understanding adhesion of microbes, both maternally provisioned and egg secreted lysozymes
(which could explain the rejection of certain microbes during the initial colonization process),
and metabolite secretion (which could explain the temporal changes in the egg associated
microbial communities). Transcriptome analysis of eggs at different time points during
embryogenesis would help to elucidate these mechanisms.
Another factor that we have not yet explored deals with fungal infection. While
conducting this experiment, we observed that when the development of eggs was arrested, eggs
became susceptible to fungal infection. These fungi can spread to other eggs and may cause
mortality of other eggs lying in the same incubation tray. It would be interesting to investigate if
any microbial species protect eggs from such fungi. In addition, it would be interesting to

198

investigate the extent to which these fungal populations affect the microbial community on the
egg surfaces via predation.
Another topic that we briefly touched on but did not explore fully is the topic of vertical
transmission of symbiont strains from maternal to offspring. We found that some microbial
communities were present on the unfertilized egg surfaces (as displayed using both TRFLP and
clone library in chapter 5). The unfertilized egg surface communities were similar to each other,
suggesting a non-random association of microbes among the unfertilized eggs. This together
with the fact that Acidovorax sp. F19 was clustered with a clade found in clone libraries of
aseptically harvested eggs suggests that Acidovorax sp. F19 may be vertically transmitted by a
female. This hypothesis can be experimentally tested in the future by sequencing microbial
genomic samples aseptically collected from different females using pyrosequencing. Using this
method, it may be possible to detect that symbiont populations have diverged within each lineage
of females if they are indeed vertically transmitted from maternal to offspring.
It would also be worthwhile to explore whether eggs from different Lake Sturgeon
populations have different relationships to microbes on their surfaces. We found that the egg
related effect was a dominant factor in shaping the microbial community on the egg surfaces, but
this egg related process could vary with different hosts and could potentially shape the egg
microbial community differently. The effect of host genetic variation on the egg microbial
communities would be a particularly interesting avenue to explore in the future. Host variation
can be considered across populations (Black lake population versus Wisconsin population)
within the same species, or across different species of sturgeon (Lake sturgeon versus white
sturgeon). This approach could even be extended to span vastly different fish species (e.g.
salmon, walleye, bass).

199

Although we focused our study on the egg associated microbial community, it would be
interesting to extend this approach to different life stages of the Lake Sturgeon. Future research
could follow the changes in microbial community assembly from eggs to larvae (including
different parts of the body such as the gut), larvae to juveniles, and juveniles to adults. We
hypothesize that there are some microbial communities already developed on the skin of larvae
immediately after hatching or even immediately before hatching (while moving inside the egg
case).
Our study on antagonistic interactions among egg isolates was conducted on a one on one
basis, but interactions among more than two microbes can occur (and are probably more likely)
within the egg surface microbial communities. It would be important in the future to investigate
the interactions among multiple isolates by co-culturing them. We also did not pursue the
correlation between antagonistic interaction and biofilm formation, but it is possible that
antagonistic interaction induces biofilm formation. These questions can be answered by
inoculating multiple strains in a broth culture and examine the relative abundance of each strain
in both broth in tubes and biofilm formed on the surfaces of tubes.
We hypothesized that Actinobacteria and Firmicutes, which occupy a significant fraction
in water microbial community, are sensitive to lysozyme on the egg surfaces. Others may able to
experimentally test this hypothesis in the future by collecting isolates of these two phyla from the
water column and testing them against a known chicken egg lysozyme or lysozyme purified from
fish eggs, if available.
The degree of adhesiveness of each isolate to the egg surface would be another
interesting avenue to take in the future. Although we only inoculated two different isolates on
eggs during fertilization, the two we chose showed totally different outcomes. One successfully

200

colonized egg surfaces, but the other did not. It would be worthwhile in the future to inoculate
other 20 egg isolates by controlling concentrations of inocula and subsequently observing the
differences in efficiency of colonization.

Implications and significance
This is the first microbial succession study conducted on fish eggs. This study
demonstrated that microbial succession can happen even within the short time span of an
incubation period. We also demonstrated that egg associated microbial communities can be
altered by many different factors, including both host and environmental factors such as stream
flow rate, temperature, and aquatic microbial community structure. The host factors are
presumably largely mediated by maternally provisioned lysozyme, which is the most dominant
factor of all and acts quickly within 6 hours post-fertilization. We also identified a putative
symbiont for eggs of the Lake Sturgeon (Acidovorax sp.). Our study provides management
implications for conserving Lake Sturgeon populations by suggesting that damming streams can
alter aquatic microbial community and temperature, which in turn can alter the microbial
communities on sturgeon eggs and life history of the sturgeon. Our study also contributes to the
broader literature on microbial community assembly and succession by demonstrating that both
local deterministic processes and dispersal play roles in shaping the microbial communities
assembly on the egg surfaces.

201