BIOTIC AND ABIOTIC FACTORS INFLUENCE FORMATION AND ONTOGENIC
DYNAMICS OF MOLECULARLY DEFINED GASTRO-INTESTINAL MICROBIAL
COMMUNITIES IN LAKE STURGEON (Acipenser fulvescens) AND CHANNEL
CATFISH (Ictalurus punctatus).
By
Shairah Abdul Razak

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Fisheries and Wildlife – Doctor of Philosophy
Ecology, Evolutionary Biology and Behavior – Dual Major
2017

ABSTRACT
BIOTIC AND ABIOTIC FACTORS INFLUENCE FORMATION AND ONTOGENIC
DYNAMICS OF MOLECULARLY DEFINED GASTRO-INTESTINAL MICROBIAL
COMMUNITIES IN LAKE STURGEON (Acipenser fulvescens) AND CHANNEL
CATFISH (Ictalurus punctatus).
By
Shairah Abdul Razak
Gastrointestinal (GI, gut) microbial communities (microbiota/ microbiomes) play
essential roles in host development and physiology. During early life stages, fish gut
microbiome composition is shaped by complex interactions of factors including dispersal
of bacteria from the surrounding water, age-dependent changes in the gut ecosystem, and
changes in dietary regimes. To investigate ecological processes that generate and
maintain compositional patterns of gut microbiome diversity, I integrated molecular
methods with experimental gut microbiome research and community ecology theory in
two important fish species, lake sturgeon (Acipenser fulvescens) and channel catfish
(Ictalurus punctatus) during early larval stages. Sequence reads of 16S rRNA gene were
analyzed using multivariate ordination methods based on Bray-Curtis distance matrices
followed by hypothesis testing using permutational multivariate analysis of variance
(PERMANOVA). In the first chapter, sturgeon larvae were raised in four rearing
hatchery treatments representing a combination of two factors: water source (stream vs
ground water) and diet (Artemia vs Artemia supplemented with detritus). As fish grew,
microbiota shifted from dominance by phyla Proteobacteria to Firmicutes. Water
possibly served as the primary bacterial inoculant during early (pre-feeding) stages of
development. Neutrality tests indicated that neutral processes were not strongly
structuring community composition. Sturgeon gut microenvironment appears to have

selected for microbial taxa, regardless of differences in treatments. The second chapter
focused on alteration of gut microbiota modulated by diet and nutrient availability.
Sturgeon gut microbiota differed among fish raised on different dietary regimes (control
vs transition). Gut microbiota of fish exposed to a dietary transition from Artemia to
frozen Chironomids were dominated by the genus Aeromonas (phylum Proteobacteria)
while fish from the control group were dominated by genus Clostridium_sensu_stricto
(phylum Firmicutes) at the end of experiment. Screening for cultured bacteria with
extracellular protease activity revealed that fish fed with Chironomids harbored proteasepositive taxa from phylogenetically distinct and more diverse clades. Next, I documented
the impact of prophylactic treatments (Chloromine-T, NaCl followed by hydrogen
peroxide; and hydrogen peroxide) on lake sturgeon larvae gut microbiota compared to
ambient environmental conditions (control) using hatchery-produced and wild-origin
fish. Gut microbiome responses to prophylactic treatments were found to be inconsistent
across fish genotypes. The last chapter documented variability in channel catfish larval
gut microbiome composition among families and nursery ponds characterized by
different rearing water in a pond aquaculture setting. I documented a strong influence of
rearing environment (pond water) on gut microbiome colonization. Gut microbiome
composition was dominated by different phyla before (Proteobacteria) vs after pond
stocking (Firmicutes). Different predominant genera were also detected over time.
Results from my research inform community ecology theory concerning effects of
stochastic and deterministic forces affecting microbial community establishment and
stability. This research is also relevant for downstream applications incorporating
microbial-based management strategies in commercial and conservation aquaculture.

Copyright by
SHAIRAH ABDUL RAZAK
2017

ACKNOWLEDGEMENTS

All praise to the Almighty for bestowing me with health, strength, and the perseverance
that helped me to finally reach the finishing line.

First and foremost, I would like to express my sincere gratitude to my advisor, Prof. Dr.
Kim Scribner for his continuous support throughout my doctoral study and dissertation
research. His invaluable knowledge, wisdom, patience, and passion in his works has been
a great source of inspiration and I aspire to be like him someday. Kim was always willing
to help me and ensured that my success in graduate school remained a top priority. Thank
you, Kim, for all your kind assistance and thoughtful advises for me, both personally and
professionally. I could not have imagined having a better advisor and mentor for my
Ph.D. study.
I would also like to thank the rest of my graduate committee: Dr. Terrence Marsh
(Terry), Dr. Mohamad Faisal, Dr. Scott Winterstein and Dr. Brian Maurer, for their
insightful comments and encouragement, but also for those challenging questions that
encouraged me to broaden my research scope to encompass these various perspectives.
To Terry and Dr. Ned Walker especially, thank you very much for providing me with
access to your laboratory and research facilities that make it possible for me to conduct
this research. I also appreciate all the knowledge and guidance that you have shared with
me. All those invaluable knowledge and experience that I gained from you will be
beneficial for my future development.
To the past and current members of Scribner Lab and ROME Lab, I am very

v

much indebted to all of you. When I came in East Lansing 5 years ago, I received
tremendous support from Jeannette Kanefsky, John Bauman, Lisette Delgado, Roshan
Angoshtari, Kari Dammerman, Hope Draheim, Britton Hildebrant, Nick Sard, Yen
Duong, Justin Waraniak, Lydia Wassink, Rob Hunter, Jonathan Hegna, and many others
that supported my Ph.D. journey here, along with your wonderful friendship. To Shaley
Valentine, I truly appreciate your effort to commit with the experiment at the Black Lake
Facilities while I was away. To the catfish research group from Mississippi State
University especially Dr. Matt Griffin, thank you for allowing me to be a part of your the
research team.
To my dear husband Khair, I could not say thank you loud enough to express how
grateful I am to always have you by my side all these while. Thank you for reassuring me
during those moments when I had doubts. Thank you for caring for our little ones when I
have to spend my nights at the laboratory finishing the works or being away for field
works. We have been through thick and thin, yet, your endless love and support is the one
true thing that keeps me going forward. To all my family members in Malaysia: my
parents, brothers, and sisters, in-laws, thank you for lending a helping hand to look after
the little ones while Khair and I are thousands of miles away from them. All your prayers
and wishes have fueled me - along my exploration. To besties, friends, and colleagues,
thank you for all your kind words and motivations. I can’t begin to tell you how much
they mean to me during the rough times. To all my fellows Malaysian here, many thanks
for being a family and making me feel at home while we were here. We definitely will be
friends for life!
Last but not least, my sincere thanks goes to all of the organizations that provided

vi

funding support for my studies and research. The Ministry of Higher Education (MOHE)
Malaysia and UKM provided financial support - in the form of a scholarship. In addition,
I would like to express my sincere thanks to The Graduate School, College of Agriculture
and Natural Resources (CANR), Department of Fisheries and Wildlife, Ecology,
Evolutionary Biology and Behavior (EEBB), and Liberty Hyde Bailey scholar program
that offered me with numerous financial awards and grants to help me complete my
studies as well as leadership opportunities that enhanced my professional development.
Thank you also to the Great Lakes Fishery Trust and Michigan Department of Natural
and Resources for funding my research.

“Not all those who wander are lost” J. R. R Tolkien

SHAIRAH ABDUL RAZAK

vii

TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................. xi
LIST OF FIGURES ......................................................................................................... xiii
GENERAL INTRODUCTION .........................................................................................1
LITERATURE CITED ........................................................................................................8
CHAPTER 1: LAKE STURGEON GUT MICROBIOTA ASSEMBLY AND
SUCCESSIONAL DYNAMICS ALLOW ECOLOGICAL EVALUATIONS OF
NEUTRAL VS HOST-SELECTIVE PROCESSES
............................................................................................................................................12
ABSTRACT .......................................................................................................................12
INTRODUCTION .............................................................................................................14
METHODOLOGY ............................................................................................................20
Experimental design and feeding regime.............................................................20
Sample collection ..................................................................................................21
DNA extraction and 16S rRNA amplicon sequencing ........................................22
Sequence processing .............................................................................................24
Analyses of bacterial community profiles and ecological statistics analyses .....24
1.1 Alpha diversity ...............................................................................................24
1.2 Temporal and Differential abundance of OTUs ..........................................25
1.3 Beta diversity ..................................................................................................25
1.4 Influence of water and diet treatments of GI tract microbiota ....................26
1.5 Tests evaluating whether GI tract communities were a neutral subset of the
environmental (source) communities ...........................................................27
RESULTS ..........................................................................................................................28
Sequencing and sample summary ........................................................................28
Ontogenetic changes of dominant bacterial taxa in lake sturgeon larvae rearing
................................................................................................................................28
Variation in gut bacterial community profiles in association with water, diet,
and time ................................................................................................................30
Neutral processes are not the dominant mechanisms generating and
maintaining community composition during early gut microbiome assembly ...32
DISCUSSION ....................................................................................................................34
APPENDIX ........................................................................................................................43
LITERATURE CITED ......................................................................................................54

viii

CHAPTER 2: COMPOSITIONAL DYNAMICS OF LAKE STURGEON GUT
MICROBIOMES ASSOCIATED WITH DIETARY TRANSITION DURING
EARLY ONTOGENETIC STAGES
............................................................................................................................................62
ABSTRACT .......................................................................................................................62
INTRODUCTION .............................................................................................................64
METHODOLOGY ............................................................................................................69
Fish husbandry and feeding experiment..............................................................69
Sample collection ..................................................................................................71
Fish dissection .......................................................................................................72
Isolation of bacterial culture & extracellular protease screening ......................72
DNA extraction and 16S rRNA amplicon sequencing ........................................74
Sequence processing .............................................................................................75
Morphometric data and survival analyses ...........................................................76
Analyses of bacterial community profiles and ecological statistics analyses .....77
2.1 Alpha diversity .................................................................................................77
2.2 Beta diversity ...................................................................................................77
Inferring gut-associated microbial communities function ..................................78
RESULTS ..........................................................................................................................80
Growth performance and survival........................................................................80
Characterization of diversity and proportion of microbial phyla in gut
community composition of lake sturgeon ............................................................81
Association between gut microbial community composition and feeding
treatments, across sampling periods…………………………………………….83
Predicted functional roles of lake sturgeon gut microbiota ................................84
Activity-based screening for protease positive isolates ........................................85
DISCUSSION ....................................................................................................................87
APPENDIX ........................................................................................................................94
LITERATURE CITED ....................................................................................................109
CHAPTER 3: CHANGES IN LAKE STURGEON GUT MICROBIOMES IN
RESPONSE TO CHEMOTHERAPEUTANT TREATMENTS
..........................................................................................................................................117
ABSTRACT .....................................................................................................................117
INTRODUCTION ...........................................................................................................118
METHODOLOGY ..........................................................................................................123
Fish husbandry ...................................................................................................123
Experimental design ...........................................................................................124
Fish dissection, DNA isolation, and PCR validation.........................................126
16S rRNA amplicon sequencing and sequence pipeline analyses ....................126
Analyses of bacterial community profiles and ecological statistics analyses ...128
3.1 Alpha diversity ...............................................................................................128
3.2 Beta diversity .................................................................................................129
3.3 Differential abundance of OTUs and biomarker identification across
treatments ............................................................................................................130

ix

RESULTS ........................................................................................................................132
Diversity of gut microbial community composition ...........................................132
Association between gut microbial community composition among fish groups
and prophylactic treatments................................................................................135
Identification of bacterial taxa influenced by chemotherapeutant treatments 136
DISCUSSION ..................................................................................................................138
APPENDIX ......................................................................................................................146
LITERATURE CITED ....................................................................................................157
CHAPTER 4: BIOTIC AND ABIOTIC FACTORS INFLUENCING CHANNEL
CATFISH EGGS AND GUT MICROBIOME DYNAMICS DURING
EARLY LIFE STAGES
..........................................................................................................................................164
ABSTRACT .....................................................................................................................164
INTRODUCTION ...........................................................................................................165
METHODOLOGY ..........................................................................................................170
Fish husbandry and spawn collection ................................................................170
Sample collection ................................................................................................171
Fish dissection .....................................................................................................172
DNA extraction and 16S rRNA amplicon sequencing ......................................172
Sequence processing ...........................................................................................173
Analyses of bacterial community profiles and ecological statistics analyses....174
4.1 Alpha diversity ...............................................................................................174
4.2 Temporal and differential abundance of OTUs...........................................175
4.3 Beta diversity .................................................................................................175
4.4 Influence of spawn, pond effects on intestinal microbiota..........................176
4.5 Contribution of predominant taxa to sample communities’ composition ...177
RESULTS ........................................................................................................................178
Pond water quality, zooplankton and phytoplankton assessment .....................178
Channel catfish gut contents and parentage determination sturgeon ..............178
Sequencing and sample summary ......................................................................178
Microbial community taxonomic composition ..................................................179
Microbial community diversity ...........................................................................180
Variation in gut bacterial community profiles in association with water and
stages....................................................................................................................180
DISCUSSION ..................................................................................................................184
APPENDIX ......................................................................................................................191
LITERATURE CITED ....................................................................................................205

x

LIST OF TABLES

Table 1.1. R2-values, p-values, and Least-square mean values of least-square mean
analyses performed on significant PCo axes for each sampling period. Five axes across
all three stages that showed significant effect of either water, food, or interaction of both
water and food treatments on the microbial community composition were Axis 6 for prefeeding, Axis 1 and Axis 5 for stage at one-week post-active feeding, Axis 1 at two-week
post-active feeding .............................................................................................................44
Table 1.2. Results of neutral models applied to gut microbiota showing the number and
proportion of shared OTUs detected in both thegut microbial communities and potential
microbial source (water). Over-represented taxa are those that were selected for (i.e.,
abundance low in water, but detected in higher abundance in the gut) whereas underrepresented taxa are present in lower abundance in the gut than in water .........................45
Table 2.1(a) Split-plot ANOVA (mixed design ANOVA) table for total length indicate
source of variability between-tank replicate associated with feeding treatment and withintank replicate associated with sampling times. A significant interaction was also observed
between treatment and stages.............................................................................................95
Table 2.1(b) Split-plot ANOVA (mixed design ANOVA) table for average fish individual
weight indicate source of variability between-tank replicate associated with feeding
treatment and within-tank replicate associated with sampling times. A significant
interaction was also observed between treatment and stages ............................................95
Table 2.2 (a) PERMANOVA analysis table indicates that variability among fish gut
microbiota differed significantly among developmental stages and between treatments. A
significant interaction was also observed between treatment and stages (PERMANOVA
test pseudo-F = 2.928, R2 = 0.087, p < 0.001; permutation=1000) ...................................96
Table 2.2 (b) Goodness of fit from linear regression model (R2), p-value, and Leastsquare means analyses performed on significantly important PCo axes calculated
separately for each stages period. PCo Axis 1 for stage 21dpf and PCo Axis 1 for stage
36dpf showed significant different across diet treatment influencing the microbial
community composition, but none of PCo Axis for stage 14 dpf was found to be
significant ...........................................................................................................................96
Table 2.3. List of taxa with taxonomic identification that show high correlation with
eigenvector of the first three Principal Coordinate axes associated with largest
eigenvalues .........................................................................................................................97

xi

Table 3.1. (a) Models generated with both fixed and random variables to explain
variation among samples of all treatments and families/group for Inverse Simpson
diversity indices. The model with the lowest AIC scores was shown in bold .................147
Table 3.1. (b) Models generated with both fixed and random variables to explain
variation among samples of all treatments and families/group for a number of observed
taxa (OTU richness). The model with the lowest AIC scores was shown in bold ..........147
Table 3.2. PERMANOVA showing variability among fish gut microbiota across all
samples. Results revealed a significant interaction of both treatments (D) and family
effect (F) influencing gut microbial communities composition for at least one samples
across treatments and families/group (PERMANOVA test pseudo-F = 1.675, R2 = 0.093,
p < 0.001; permutation=1000) .........................................................................................148
Table 3.3. Goodness of fit from linear regression model (R2), p-value, and Least-square
means analyses performed on significantly important PCoA axes calculated separately for
each family/group. (a) PCoA Axis 2 associated with second largest eigenvalue for fish
from family HA show that microbial community composition in control group
significantly differed than other treatments; (b) PCoA Axis 6 associated with sixth largest
eigenvalue for fish from group WB show that microbial community composition in both
control and salt fish significantly influence by each respective treatment ......................148
Table 4.1. PERMANOVA analysis indicating lack of significant differences between
centroid location of egg microbial communities from different brood ponds
(PERMANOVA test pseudo-F = 1.258, R2 = 0.386, p = 0.333; permutation=24*)........192
Table 4.2. PERMANOVA analysis indicating no significant differences between centroid
location of gut microbial communities were observed between fish from swim-up and
stocking developmental stages (D). Family (F: 147 A,147 B, 154 A, and 154 B) were
nested within brood pond (B: 147 vs 154) .......................................................................192
Table 4.3. PERMANOVA analysis indicating summarizing partitioning of variability
among fish gut microbiota from samples collected during 24 hr and 21 d post-release into
two different nursery ponds significantly influenced by developmental stages (D), family
(F), and nursery pond (P). Results revealed a significant interaction of these factors
influencing the microbial communities present in the gut (PERMANOVA test pseudo-F
= 1.863, R2 = 0.038, p < 0.05; permutation =1000) .........................................................193

xii

LIST OF FIGURES

Figure 1.1. Bacterial composition of different communities identified from (a) lake
sturgeon larval gut and (b) environmental samples ...........................................................46
Figure 1.2. Estimates of Alpha diversity for lake sturgeon gut microbial communities
from all treatments across all developmental stages ..........................................................48
Figure 1.3. Visual representation of ontogenetic changes in lake sturgeon larval gut
microbiota based on Bray-Curtis distances. Bray-Curtis dissimilarities between
communities originating from gut, water and food samples from all four different
treatments at three developmental stages (pre-feeding, a week active feeding, two week
active feeding) were visualized using Principal coordinates analyses (PCoA, or also
known as Metric Dimensional Scaling MDS) plots. .........................................................50
Figure 1.4. Interaction plots of marginal (least-square, LS) means for first PCoA axes
(axes that explained largest variation in dataset) at different developmental stages. ........51
Figure 1.5 Results of neutral model testing with water as the source of gut microbial
communities (a) at pre-feeding stage, (b) after one-week active feeding, (c) after two
weeks of active feeding ......................................................................................................53
Figure 2.1. Mean proportion of cumulative survival (solid-line) of lake sturgeon from
Transition (TR) and Control (CR) treatment group with 95% CI (dashed-line) ...............98
Figure 2.2. Morphometric data with error bar representing standard error mean (a) total
length (mm) by treatment group, at each sampling point; (b) average weight (g) of fish
from each treatment group fed with control food vs fish transitioned to Chironomid
larvae. * indicate significant differences in treatment mean test based on Welch-t test (*
p-value < 0.05; ** p-value < 0.01; *** p-value < 0.001) ..................................................99
Figure 2.3 (a) Relative abundance (percentage) of six bacterial phyla (Acidobacteria,
Actinobacteria, Bacteroidetes, Firmicutes, Fusobacteria, Proteobacteria) found in gut
microbiota of lake sturgeon larvae for each treatment at different times. The remaining
taxa were assigned as Others. (b) Relative abundance (percentage) of six bacterial phyla
(Acidobacteria, Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria,
Verrucomicrobia) found in aquatic samples. The remaining taxa were assigned as Others.
..........................................................................................................................................101

xiii

Figure 2.4 (a) Inverse Simpson Diversity index and (b) number of observed OTUs/ Taxa
richness for both feeding treatment (Control, CR and Transition, TR) at different
sampling time ...................................................................................................................102
Figure 2.5. Visualization using Multivariate Principal Coordinates Analyses (PCoA) plots
of variation in lake sturgeon gut and water microbial community composition among
collections made at different times and between food from different treatment groups
using Bray-Curtis distances. Samples were taken from replicates of different treatments at
three developmental stages (before transition at 14dpf, during transition week at 21dpf,
after transition at 36dpf). (a) Plot based on Bray Curtis distance Axis 1 and 2, (b) Plot
based on Bray Curtis distance Axis 1 and 3, (c) Plot based on Bray Curtis distance Axis 2
and 3 ..........................................................................................................................103,104
Figure 2.6. Relative abundance of top 20 annotated functional groups identified using
KEGG Orthologs pathway categorized at molecular-level 2 associated with lake
sturgeon larval microbiome from two diet treatments, Control (CR) and Transition
(TR) at each sampling point. (a) 14 dpf; (b) 21 dpf; (c) 36 dpf .......................... 105,106
Figure 2.7 Weighted Neighbor-Joining tree showing evolutionary relationship for
selected top match taxa with single outgroup species. Branch supports were estimated
using bootstrap analysis based on 100 pseudoreplicates for NJ and the percentage of
bootstrap value were displayed next to branches in the inferred phylogeny. (a) NJ tree
for isolates from fish in the CR group; (b) NJ tree for isolates from fish in the TR
group .......................................................................................................................... 107
Figure 3.1. Schematic design of the larval chemotherapeutant study .............................149
Figure 3.2 Taxonomic composition of bacterial communities identified from the lake
sturgeon larval GI tracts (a) at the phyla level and (b) at the genera level ..............150, 151
Figure 3.3 Estimates of alpha diversity (a) Inverse Simpson index; (b) number of
observed taxa (OTU richness) for lake sturgeon gut microbial communities from all
samples of treatments and families/groups. Each bar indicates mean with S.E. for each
treatment from each family/group ...................................................................................152
Figure 3.4. Visual representation of differences in the gut microbiota of lake sturgeon
larvae (Bray-Curtis distances). Bray-Curtis dissimilarity distances between communities
from fish exposed to four chemotherapeutant across families/group were visualized by
Principal Coordinates Analysis (PCoA) plots..................................................................154
Figure 3.5. Interaction plot of marginal (least-square, LS) means for significant PCoA
axes detected from hatchery family (HA) and wild group B (WB). Axes represent
variation in taxonomic diversity and relative abundance for a given family/group. (a) A
significant difference exists between microbial communities composition of HA fish in
control group compared to fish that received other prophylactic treatments; (b)
Significant difference exists in the gut microbiota of fish from WB associated with

xiv

different prophylactic treatments. See details in Table 3.3 ..............................................155
Figure 3.6. LEfSe-detected taxa from one-to-all comparison (control to all other three
chemotherapeutant). .........................................................................................................156
Figure 4.1 Schematic design of the larval catfish microbiome study ..............................194
Figure 4.2 Phyla level composition of bacterial communities identified from channel
catfish fry GI tracts (a) prior to stocking into nursery ponds (NP) and (b) post stocking
into nursery ponds ............................................................................................................195
Figure 4.3 Phyla level compositions of bacterial communities identified from water
samples (a) prior to fish stocking into nursery ponds (NP) and (b) after fish stocking into
nursery ponds ...................................................................................................................197
Figure 4.4. Estimates of Alpha diversity (Inverse Simpson index, (a); OTU richness
(Number of observed taxa), (b)) for channel catfish gut microbial communities from all
treatments across all developmental stages......................................................................199
Figure 4.5. Visual representation of differences in gut microbiota of channel catfish fry
(Bray-Curtis distances). Bray-Curtis dissimilarities distance between communities
originating from fish GI tracts and water samples for all developmental stages examined
from both spawns are visualized using Principal coordinates analyses (PCoA) plots.
Points represent a bacterial community from each sample (a) Two dimensional
representation of PCoA using first and second axis; (b) first and third axis; (c) second and
third axis...................................................................................................................201, 202
Figure 4.6 Genus-level contribution to gut microbiome community variation. Heat maps
show the relative abundance value of each bacterial genus present in gut microbiota for
each sample for all developmental stages. These values are depicted by color intensity
with the legend indicated on the top of the figure ...........................................................203

xv

GENERAL INTRODUCTION
The presence of compositionally diverse microbial communities (also known as
microbiota / microbiome) inside gastro-intestinal (GI) tracts has important consequences
to proper physiological functioning of the animal hosts. The consortium of
microorganisms play crucial roles in many host physiological functions including, but not
limited to, nutrition and metabolism functions, development and somatic growth,
immunity, and protection against pathogens (Bjorksten, 2006; Hattori & Taylor, 2009;
Robinson et al., 2010; Sekirov et al., 2010). Recent expansion and advancement in
sequencing technology, in particular the utilization of molecular tools such as
metagenomic analyses based on 16S rRNA genes, has led to a growing literature that has
characterized the structure and function of the gut microbiota for many taxonomic groups
including fish (Ghanbari, Kneifel, & Domig, 2015).
Fish are an important vertebrate group that comprises the vast majority of
vertebrate species (nearly 30,000). Taxa show remarkable variation in living
environments, ecology, and evolutionary characteristics (Nelson, 2016). Studying gut
microbiota in fish is warranted because of the extensive manipulation of microbial
communities in aquaculture industries that crucial for food production and economic
growth. Interactions of gut microbiota with fish hosts have profound impacts (Nayak,
2010; Romero, Ringø, & Merrifield, 2014; Vadstein et al., 2012), as fish are constantly
immersed in aquatic solutions that have high bacterial load (De Schryver & Vadstein,
2014; Skjermo & Vadstein, 1999; Vadstein et al., 2012; Verschuere et al., 2000). In
natural and aquaculture settings, wild and cultured fish typically experience a high level
of mortality, particularly during early life stages in part due to interactions between the

1

microbes from surrounding environment and microbes associated with larval fish itself
(Hamre et al., 2013; Oliva-Teles, 2012; Vadstein et al., 2012). For threatened fish species
like the lake sturgeon (Acipenser fulvescens) and important fish for food production like
the channel catfish (Ictalurus punctatus), the high mortality rates experienced during
early ontogenetic stages have been identified as likely bottlenecks to successful
recruitment for adult fish populations (Caroffino et al., 2010; Forsythe et al., 2013;
Hargreaves & Tomasso, 2004; Hawke & Khoo, 2004).
Colonization of fish GI tracts by microbes occurs during early life stages and
these microbial communities can later give rise to relatively stable communities in an
adult fish. Yet, extensive spatial, temporal, and inter-individual variation does exist in the
composition of gut microbiota which play important roles in homeostatic regulation
(Costello et al., 2012; Llewellyn et al., 2014; Yatsunenko et al., 2012; Wong et al., 2015).
Ecological processes that underlie the formation and temporal/ontogenetic dynamics of
gut microbiota composition are poorly understood (Adair & Douglas, 2017; Christian et
al., 2015). Little research has been conducted on how these processes interact in fish
hosts (Llewellyn et al., 2014). From an ecological point of view, the GI tract and the
microbiota residing within this anatomical structure can be viewed as a distinct
ecosystem with its own community. Thus, the general field of community ecology can
provide a robust conceptual basis to understand the pattern and diversity of gut
microbiota (Adair & Douglas, 2017; Christian et al., 2015; Costello et al., 2012; Vellend,
2010; Zeng et al., 2015).

2

Due to the importance of gut microbiota to fish hosts, and the complex interaction
of fish, microbes, and environmental factors, it is important to investigate the interrelationships between the fish hosts and their surrounding environment (rearing water and
diets) and how these affect microbial community composition and diversity in the fish
gut during early life stages. Using gut microbiota as an empirical system, the theoretical
foundation and conceptual understanding of processes that generate and maintain
compositional patterns and taxonomic gut microbiota diversity will provide insights into
the relative importance of ecological and evolutionary processes.
My dissertation is composed of four chapters. In the first chapter, experimental
manipulation of gut microbiota colonization in lake sturgeon larvae was performed using
full-sib fish produced from a single spawning pair that were raised in environments that
differed in the water source (stream vs filtered ground-water) and diet (supplemented vs
non-supplemented Artemia). Using massively parallel sequencing of a portion of the 16S
rRNA gene from the GI tracts of larval lake sturgeon, we quantified the gut microbial
composition and taxonomic diversity at three stages (pre-feeding, one, and two weeks
after active feeding began). This research addressed two main questions: (1) What
factor(s) and/or ecological processes affect the variation in microbial community
composition that colonized the GI tracts of lake sturgeon larvae during early ontogenetic
stages? (2) How does microbial community composition and diversity vary through early
ontogenetic stages as the fish host develops and transitions in different environments? I
hypothesized that the colonization and development of gut microbial communities would
be influenced by interactions between factors including host ontogenetic stage, diet and
rearing environment. Findings showed that the composition of microbial community

3

varied as the fish ages. Data further suggest that water served as the primary inoculant of
the gut communities at early stages of development. As larvae began to feed
exogenously, the intestinal habitat of sturgeon guts appears to have selected for specific
microbial taxa, regardless of the differences in host rearing environment and feeding
regime. Neutral expectation (i.e dispersal) underlying the colonization of these gut
microbiota was supported statistically but was not a dominant process.
The second chapter focused on specifically, the taxonomic diversity and gut
microbiota composition at the early ontogenetic stages as the consequences for fish with
compromised nutritional status associated with dietary transition. Our objectives were to:
(i) document the effect of food-transition on the gut microbial composition, (ii) screen
and identify the presence of protease-producing bacteria in fish from different feeding
treatments. Lake sturgeon that experienced diet switch from brine shrimp to frozen
bloodworms performed poorly in terms of growth and survival. Prior to the experiment,
fish were expected to perform better as they successfully transitioned to newly offered
prey. Results could be due to starvation in this group relative to the control group that
was only fed brine shrimp throughout the experiment. Compositional changes were
observed throughout the fish developmental stages, but fish in transitioned group
exhibited distinct communities compared to the fish in control group at the end of
experimental duration. We also observed higher taxonomix diversity of proteaseproducers in the GI tract of fish, which received bloodworms (transitioned group).
The third chapter investigated the influence of prophylactic treatments using three
chemotherapeutants: (i) immersion in antibiotic Chloramine-T, (ii) immersion in salt
followed by hydrogen peroxide after 24 hr, and (iii) immersion in hydrogen peroxide; on

4

the taxonomic composition and diversity of gut microbiome in the lake sturgeon larvae in
comparison to a control (no treatment) group. Larvae from two hatchery-produced
families and from wild eggs collected at two natural spawning sites in the Black River
(Michigan) were exposed to weekly prophylactic treatments. I hypothesized that
disturbances in their gut microbiota caused by different treatments would decrease
diversity and would increase community compositional differences in taxonomic
composition. As fish are treated with chemotherapeutants, they are believed to experience
unfavorable rearing conditions in aquaculture settings that could interfere with their
physiology. However, in human and other vertebrates, chemicals might be harmful to
some microbes and may affect the composition and stability of their gut microbiomes.
We asked the following questions: (1) what impact will the prophylactic chemical
treatments (salt (NaCl), antibiotics (Chloromine-T), and hydrogen peroxide) that are
commonly used in fish aquaculture have on the gut microbiome of lake sturgeon? (2) do
microbiota of fish exposed to prophylactic treatments differ from fish in the control
treatment or do communities show resistance following treatment disturbance? Statistical
tests of data generated from 16S rRNA gene sequencing indicated that variation in alpha
diversity indices of gut microbiomes were greater in fish samples from different origins
(hatchery families vs wild caught eggs) relative to variation among fish exposed to
different treatments. In one hatchery-produced family, gut communities of fish from the
control group were distinct relative to gut communities of fish exposed to all three
chemical treatments.

5

My final dissertation chapter emphasized common themes in microbial
community ecology. In this chapter, we focused on commonalities and differences in
fish-host microbiome associations in the channel catfish. In contrast to lake sturgeon, the
channel catfish is a warm water species, characterized by a different feeding ecology and
occupancy of different environments (pond aquaculture) during different ontogenetic
stages, which we hypothesized, would influence their gut bacterial community
composition and diversity. This study addressed the following questions: (i) what factors
influence the formation of channel catfish gut microbiomes when raised in pond
aquaculture following common industry practices? (ii) is there variability in microbiome
composition across families and ponds characterized by different fish rearing water
environment? This chapter involved collaboration with a research group from Mississippi
State University (MSU) as part of an effort to identify best management practices to be
adopted for catfish rearing in order to combat pathogen outbreaks. Results indicated
strong selection occurred within the host where specific taxa proliferated and appeared to
have been favored. The gut community also exhibited some resemblance to
corresponding water samples. Community characterizations revealed temporal shifts in
community composition and the influence of microbes present in the rearing water
supply. Implications of modes of gut colonization to proposed probiotic uses in catfish
pond aquaculture are discussed.
Results from my research offer empirical evaluation the effects of stochastic and
deterministic forces affecting microbial community establishment and stability, which are
important aspects community ecology theory. My research is also relevant for
downstream application in commercial and conservation aquaculture by explaining

6

factors that may impact efforts to incorporate microbial-based management strategies
such as probiotics usage.

7

LITERATURE CITED

8

LITERATURE CITED
Adair, K. L., & Douglas, A. E. (2017). Making a microbiome: the many determinants of
host-associated microbial community composition. Current Opinion in
Microbiology, 35: 23–29
Bjorksten, B. (2006). Clinical and The gut microbiota : a complex ecosystem. Clinical
and Experimental Allergy, 36: 1215–1217.
Caroffino, D. C., Sutton, T. M., Elliott, R. F., Donofrio, M. C., Utton, T. R. M. S., &
Onofrio, M. I. C. D. (2010). Early Life Stage Mortality Rates of Lake Sturgeon in
the Peshtigo River, Wisconsin. North American Journal of Fisheries Management,
30(1): 295–304
Christian, N., Whitaker, B. K., & Clay, K. (2015). Microbiomes: unifying animal and
plant systems through the lens of community ecology theory. Frontiers in
Microbiology, 6(869): 1-19
Costello, E. K., Stagaman, K., Dethlefsen, L., Bohannan, B. J. M., & Relman, D. A.
(2012). The Application of Ecological Theory Toward an Understanding of the
Human Microbiome. Science, 336(6086): 1255–1262.
De Schryver, P., & Vadstein, O. (2014). Ecological theory as a foundation to control
pathogenic invasion in aquaculture. The ISME Journal, 8: 2360-2368.
Forsythe, P. S., Scribner, K. T., Crossman, J. A., Ragavendran, A., & Baker, E. A.
(2013). Experimental Assessment of the Magnitude and Sources of Lake Sturgeon
Egg Mortality. Transactions of the American Fisheries Society, 142(4): 1005–1011.
Ghanbari, M., Kneifel, W., & Domig, K. J. (2015). A new view of the fish gut
microbiome: Advances from next-generation sequencing. Aquaculture, 448: 464–
475.
Hamre, K., Yufera, M., Ronnestad, I., Boglione, C., Conceicao, L. E. C., & Izquierdo, M.
(2013). Fish larval nutrition and feed formulation: Knowledge gaps and bottlenecks
for advances in larval rearing. Reviews in Aquaculture, 5:S26-S58.
Hargreaves, J. A., & Tomasso, J. R. (2004). Environmental Biology. In Tucker, C. S. &
Hargreaves, J. A. (1st Eds), Biology and Culture of Channel Catfish, 34: 36–68.
Elsevier Science, Amsterdam, The Netherlands.
Hattori, M., & Taylor, T. D. (2009). The human intestinal microbiome: a new frontier of
human biology. DNA Research : An International Journal for Rapid Publication of

9

Reports on Genes and Genomes, 16(1): 1–12.
Hawke, J. P., & Khoo, L. H. (2004). Infectious Diseases. In Tucker, C. S. & Hargreaves,
J. A. (1st Eds), Biology and Culture of Channel Catfish, 34: 36–68. Elsevier Science,
Amsterdam, The Netherlands.
Llewellyn, M. S., Boutin, S., Hoseinifar, S. H., & Derome, N. (2014). Teleost
microbiomes: The state of the art in their characterization, manipulation and
importance in aquaculture and fisheries. Frontiers in Microbiology, 5 (207):1-17.
Nayak, S. K. (2010). Role of gastrointestinal microbiota in fish. Aquaculture Research,
41(11): 1553–1573.
Oliva-Teles, A. (2012). Nutrition and health of aquaculture fish. Journal of Fish
Diseases, 35(2): 83–108.
Robinson, C. J., Bohannan, B. J. M., & Young, V. B. (2010). From structure to function:
the ecology of host-associated microbial communities. Microbiology and Molecular
Biology Reviews : MMBR, 74(3): 453–476.
Romero, J., Ringø, E., & Merrifield, D. L. (2014). The Gut Microbiota of Fish. In
Merrifield, D. & Ringo, E., editors Aquaculture Nutrition: Gut Health, Probiotics
and Prebiotics, 75-100, John Wiley & Sons, Ltd., Chichester, UK.
Sekirov, I., Russell, S., & Antunes, L. (2010). Gut microbiota in health and disease.
Physiological Reviews, 90(3), 859–904.
Skjermo, J., & Vadstein, O. (1999). Techniques for microbial control in the intensive
rearing of marine larvae. Aquaculture, 177(1–4): 333–343.
Vadstein, O., Bergh, Ø., Gatesoupe, F.-J., Galindo-Villegas, J., Mulero, V., Picchietti, S.,
Scapigliati, G., Makridis, P., Olsen, Y., Dierckens, K., Defoirdt, T., Boon, N., De
Schryver, P., Bossier, P. (2012). Microbiology and immunology of fish larvae.
Reviews in Aquaculture, 5, S1–S25.
Vellend, M. (2010). Conceptual synthesis in community ecology. The Quarterly Review
of Biology, 85(2):183–206.
Verschuere, L., Rombaut, G., Sorgeloos, P., & Verstraete, W. (2000). Probiotic Bacteria
as Biological Control Agents in Aquaculture, 64(4): 655–671.
Wong, S., Stephens, W.Z, Burns, A. R., Stagaman, K., David, L. A., Bohannan, B. J. M.,
Guillemin, K., Rawls, J. F. (2015). Ontogenetic differences in dietary fat influence
microbiota assembly in the zebrafish gut. mBio, 6(5), 1–9.
Yatsunenko, T., Rey, F. E., Manary, M. J., Trehan, I., Dominguez-Bello, M. G.,

10

Contreras, M., Magris, M., Hidalgo, G., Baldassano, R. N., Anokhin, A. P., Heath,
A.C., Warner, B., Reeder, J., Kuczynski, J., Caporaso, J. G., Lozupone, C. A.,
Lauber, C., Clemente, J. C., Knights, D., Knights, R.,Gordon, J. I. (2012). Human
gut microbiome viewed across age and geography. Nature, 486(7402):222-227
Zeng, Q., Sukumaran, J., Wu, S., & Rodrigo, A. (2015). Neutral Models of Microbiome
Evolution. PLOS Computational Biology, 11(7), e1004365.

11

CHAPTER 1: LAKE STURGEON GUT MICROBIOTA ASSEMBLY AND
SUCCESSIONAL DYNAMICS ALLOW ECOLOGICAL EVALUATIONS OF
NEUTRAL VS HOST-SELECTIVE PROCESSES
ABSTRACT
Gastrointestinal (GI) or gut microbiota play essential roles in host development and
physiology. These roles are influenced in part by microbial community composition.
During early developmental stages, ecological processes underlying the assembly and
successional changes in host GI community composition are influenced by numerous
factors including dispersal from the surrounding environment, age-dependent changes in
the gut environment, and changes in dietary regimes. However, the relative importance of
these ecological processes to gut microbial communities are not well understood. We
examined effects of environmental and host factors (change in ontogenetic stages and
feeding physiology) on compositional changes in gut microbial communities based on
massively parallel sequencing of a portion of the 16S rRNA gene from GI tracts of a
primitive teleost fish, the lake sturgeon (Acipenser fulvescens). Lake sturgeon larvae
during early ontogenetic stages were raised in environments that differed in water source
(stream vs filtered groundwater) and diet (supplemented vs non-supplemented Artemia).
We quantified gut microbial composition and taxonomic diversity at three stages (prefeeding, one, and two weeks after exogenous feeding began). We documented taxonomic
compositional divergence between the gut and environmental microbial communities
among stages across each water and dietary treatment. Gut microbial community
diversity declined and community composition differed significantly among stages. Large
percentages of taxa present in the gut were over or under-represented relative to neutral
expectations in each sampling period. Findings indicate strong, dynamic relationships

12

between gut microbiota composition and host gastrointestinal physiology, with
comparatively smaller influences associated with rearing environments. Neutral models
of community assembly and compositional change could not be rejected, but selectivity
associated with microbe-host interactions were evident. Our results have implications for
lake sturgeon conservation and aquaculture production specifically, and applications of
microbial-based management in teleost fishes generally.

13

INTRODUCTION
One of the primary goals in community ecology is to understand species compositional
diversity and the relative influences of forces underlying patterns of distribution and
abundance across spatial and temporal scales (Vellend, 2010). Microbial communities,
particularly those associated with the gastrointestinal tract (GI) of vertebrate hosts have
coevolved in host-microbial ecosystems to be taxonomically and functionally diverse.
Communities have important functions associated with host nutritional, developmental,
immunological, and physiological process (Llewellyn et al. 2014).
Host gastrointestinal (GI) tracts are colonized during early ontogenetic stages by
subsets of taxa found in the environment. Microbes that occur in the gut during early life
stages can exhibit large variation in community composition (Costello, Stagaman,
Dethlefsen, Bohannan, & Relman, 2012; Llewellyn, Boutin, Hoseinifar, & Derome,
2014; Yatsunenko et al., 2012). The gut microbiota of an adult may therefore in part
reflect the history of past exposure to microbes and other past environmental factors
present during earlier ontogenetic stages. Alterations of these early life stage
communities have important consequences during later stages (i.e., ontogenetic
contingency (Diggle 1994; Orizaola et al. 2010; Huey et al. 2012), including disease
susceptibility (Gensollen et al. 2016).
Microbial community interactions with fish hosts have profound impacts because
of the of high bacterial abundance in aquatic realms (De Schryver & Vadstein, 2014;
Vadstein et al., 2013; Verschuere, Rombaut, Sorgeloos, & Verstraete, 2000). Fish rely
heavily on gut microbiota for many physiological functions including but not limited to
growth, digestion, nutrient production and absorption, protection of animal against

14

pathogens and regulation of the immune system (Llewellyn et al., 2014; Wu & Wu, 2012;
Hooper et al., 2012; Cain & Swan, 2010; Gómez & Balcázar, 2008; Macpherson &
Harris, 2004). The initial colonization of microbial communities during early life stages
can be influenced by many factors including fish host physiology, nutritional, and
surrounding environmental factors (Giatsis et al., 2014; Ingerslev et al., 2014; Llewellyn
et al., 2014; Nayak, 2010; Romero, Ringø, & Merrifield, 2014; Stephens et al., 2015).
Although the importance of the gut microbiota to fishes has long been recognized,
the compositions of these microbial communities are often highly variable across
developmental stages. Therefore researchers are challenged to infer the functional
importance of constituent taxa. Understanding processes governing community
formation, diversity, and their dynamic compositional changes remain elusive.
Understanding the principles associated with bacterial colonization of GI tracts and with
compositional changes across life stages will help managers, particularly in aquaculture
settings, to manipulate gut communities to promote animal health, performance, and
productivity.
Advances in next-generation sequencing technologies have expanded capabilities
for the study of complex intestinal microbial community in terms of composition and
functions beyond what has historically possible using culture-based methods (Dave et al.,
2012; Ghanbari, Kneifel, & Domig, 2015). Microbiomes refer to assemblages of
microorganisms existing in or associated with a defined habitats (in this case the
gastrointestinal tract); including active and interacting members as well transient or
inactive members (Lederberg & McCray, 2001; Shade & Handelsman, 2012). Research
on fish gut microbiomes has lagged behind research on other vertebrate hosts until Rawls

15

and colleagues (2004) characterized the zebrafish gut microbiome and documented the
difference between mammalian and teleost microbiota based on gnotobiotic teleost
models. Additional studies followed on fish using the same model organism (Burns et al.
2016; Stephens et al., 2016; Yan, van der Gast, & Yu, 2012, Rawls et al. 2011), while a
growing gut microbiome literature focused on other important aquaculture species like
salmonids, tilapia, carp, and others. (see review in Llewellyn et al. 2014).
The number of studies concerning fish gut microbiota has increased, however,
most involved descriptions of microbial community composition. Comparatively less
attention has been devoted to studies of ecological processes governing the composition
of fish gut microbiota. Examination of host GI tracts and microbiota residing within this
anatomical structure and ecosystem would profit from using an ecological conceptual
context (Costello et al. 2012). For example, Vellend (2010) describes four processes:
selection, drift, dispersal, and speciation that underlie patterns in ecological communities
that are also applicable to studies of microbial communities (Nemergut et al. 2013,
Costello et al. 2012). In our examination of processes affecting fish GI tract microbiomes,
we evaluate alterative predictions from neutrality theory (Hubbell 2001) with niche-based
hypotheses (Vellend, 2010).
Gut microbiota assembly typically occurs early in host development, involving
factors such as rearing conditions and diet, providing opportunities to study microbial
community formation and succession, including exchanges between the environment and
gut (Wong et al. 2016). For example, Sloan et al. (2006) applied neutral models to
prokaryotic systems. The Sloan Neutral Prokaryotic Models were calibrated using data
commonly used by microbial ecologists such as analyses of 16S rRNA gene sequences or

16

functional gene sequences. The ecological neutrality theory emphasizes the absence of
differences in per-capita growth rate, death, and dispersal among species, and assumes
equal fitness across species (Hubbell, 2001, Sloan et al. 2006, Vellend 2010, Burns et al.
2016). Deterministic processes include ontogenetic shifts in the gut environment that
place certain taxa at a selective advantage over others. Therefore, the pattern of
community composition and taxonomic diversity likely are the outcomes of stochastic
processes of dispersal and drift (Burns et al. 2016, Venkataraman et al. 2015). By
incorporating these conceptual frameworks together with advanced sequencing
technologies, an improved understanding of processes governing fish gut microbiota
assembly during early life stages is possible and warranted.
Sturgeons belong to one of the oldest groups of the bony fishes (Osteichthyes) and
many are species of conservation concern. Sturgeons are also an important group in
production and conservation aquaculture. Sturgeon species possess a unique valvular
hindgut called a spiral valve that is absent in other fish except for Elasmobranchs
(Buddington & Christofferson, 1985). There have been few studies on the gut microbiota
of these primitive fishes. The majority of studies conducted have presented data derived
from cultivable intestinal microbes on other sturgeon species (Askarian, Kousha, &
Ringø, 2009; Callman & Macy, 1984; Masouleh, Sharifpour, & Arani, 2006; Ghanbari et
al. 2009; Geraylou et al. 2012). Of all 27 sturgeon species, lake sturgeon (Acipenser
fulvescens) is only sturgeon species endemic to the Great Lakes Basin and the only
sturgeon in the genus Acipenser that spends their life solely in freshwater
(potamodromous) in North America (Detlaf, Ginsburg, & Schmanlhausen, 1993; Wilson
& Mckinley, 2004). This species has experienced significant declines in abundance and

17

distribution due to overfishing and loss and degradation of the habitat (Saffron, 2004). In
recent years, sturgeon conservation aquaculture has increased greatly as part of
restoration actions.
Successful lake sturgeon production can be limited due to high mortality, especially
during early ontogenetic stages. Low survival of larvae is tied to nutritional regimes
associated with diet formulation, feeding schedule, food presentation and preference
(Mim et al., 2002; Vedrasco et al., 2002). After hatch, yolk-sac larvae gradually develop
gastrointestinal (GI) tracts that will eventually resemble adult structures by 10-11 days
post hatch (dph). Once the yolk-sac period has completed and the mouth begins to open,
fish transition to exogenous feeding, often on brine shrimp (Artemia) (Buddington &
Christofferson, 1985; Wang, Binkowski, & Doroshov, 1985). During this stage, the early
exposure and interaction of the fish gut with microbial colonists from surrounding water
is possible, and shifts in intestinal microbiota can be documented as fish continue to
develop. Early ontogenetic changes in host diet and physiology can shape gut community
dynamics.
Our objectives were to characterize gut microbiota in lake sturgeon during larval
stages at the time before they start feeding until 14 days following the onset of exogenous
feeding using Illumina Miseq high-throughput sequencing of a portion of the 16S rRNA
gene. To our knowledge, this research represents the first experimental evaluation of the
ontogenetic changes in community composition and diversity of the intestinal microbiota
in sturgeon. This study was also designed to quantify associations of host factors, water
supply, and diet on the lake sturgeon gut microbiota. We also quantify whether gut
microbiome assembly is consistent with neutral expectations. Our findings have

18

implications for the management of nutrition, disease, and potential probiotic use in lake
sturgeon culture, and characterize dynamic relationships between host ontogeny and
environmental epibiota associated with temporal variability of microbiota residing in the
gut.

19

METHODOLOGY
Experimental design and feeding regime
Lake sturgeon larvae were produced from a single mated pair collected during the
sturgeon-spawning season on the Upper Black River, Cheboygan, MI in May 2013. Fullsib individuals were used to reducing potential variability in microbiota associated with
host genetic background. All individuals were raised under four different rearing
conditions. We used a 2x2 treatment factor design associated with water type and feeding
regime. Water types included river water from the natal stream and Ultra-violet (UV)
treated ground water (GW), reflecting water sources commonly used in traditional
hatchery operations. These water types were used throughout the culture process
including food (brine shrimp or Artemia) preparation. Fish were either fed live Artemia
nauplii that are commonly used in sturgeon hatcheries (Bauman et al. 2015), or Artemia
supplemented with organic retentate including detritus and aquatic zooplankton obtained
from serial filtration through 100 µ and 50 μ filters used to filter river water entering the
hatchery (hereafter referred to as “retentate”). Presence of digestible taxa in retentate was
confirmed using massively parallel sequencing of the v9 region of the 18S rRNA gene.
All four treatment groups were denoted as S (fish raised in natal stream water, fed live
Artemia); Sp (fish raised in natal stream water, fed live Artemia fish mixed with
retentate); GW (fish raised in groundwater, fed live Artemia); GWp (fish raised in
groundwater, fed live Artemia mixed with retentate). These acronyms will be used
throughout the paper.
Each treatment included six 3.0 L polycarbonate tanks (Aquatic Habitat) that
served as biological replicates, each holding 70 individuals to achieve a statistical power

20

of 0.8. The power analyses were performed based on our preliminary microbial studies
(Fujimoto et al., 2013). Fish were exposed to the same water type beginning at the time
of egg fertilization and incubation stages. A newly hatched larval group of 70 fish were
then distributed into each treatment replicates at 10 days post hatch (dph). The food was
offered at 12 dph. However, only at 16 dph, we began to consistently offer food when at
least half of the fish were feeding. To ensure fish received consistent amounts of food
throughout the experiment, previously established dry-weight feeding rates for sturgeon
(Deng et al. 2003) were utilized whereby larvae in all tanks were fed at 26% body weight
daily (BWD). Prior to the first feeding each day, retentate were added to the freshly
prepared Artemia for treatment Sp and GWp. The amount of retentate (in grams)
collected daily varied depending on each day collection, yet the food amount was
adjusted accordingly to ensure fish in each treatment consumed the same amount of food.
Fish were fed to satiation three times daily. Mortality was recorded daily, and the body
weight (g) was recorded once in every third day. All experiments were conducted at
Sturgeon Streamside Rearing Facility managed by Michigan Department of Natural
Resources (MDNR) and Michigan State University (MSU) at Onaway, MI using
approved Institutional Animal Care and Use Committee (IACUC) protocols.
Sample collection
We sampled ten sturgeon larvae from each replicate for each treatment at three
ontogenetic developmental stages: Time 1 - before active feeding (11-day post hatch,
dph), Time 2- after one week of fish actively feeding (22-dph), and Time 3 - after two
weeks of active feeding (29-dph). These time points were selected to capture critical
phases of GI tract development after fish completely absorbed their yolk sac and

21

gastrointestinal tract anatomy was completed (approximately at10 dph). Lake sturgeon
started actively feeding between 13-16 dph when food was first offered (Buddington &
Christoferson, 1985; Buddington & Doroshov, 1986). Fish were sampled for microbial
community interrogation at each time period and were euthanized with an overdose of
MS-222 (Sigma-Aldrich, St Louis, MO, USA) at the time of sampling. Each individual
larvae was photographed at the time of euthanization and images were analyzed using
Image J software to determine total body length (mm). Fish were then transferred and
preserved in 50 ml Corning ® centrifuge tubes containing 80% filtered-sterilized ethanol
until GI tract dissection and bacterial DNA extraction was performed.
Other environmental samples including water and food were also collected during
each sampling period. 250 ml of water samples from stream water and ground water were
collected from the hatchery reservoir tank and being filtered through 0.22 um 47 mm
filter membrane (Sterlitech®) using a hand-pump to obtain the aquatic microbial
communities on 0.45 micron filter paper. These filters were then transferred and
preserved in 50 ml Corning ® centrifuge tubes containing 80% ethanol until bacterial
DNA extraction is performed. For food samples, approximately 200μl of food were
pipetted into 2ml Eppendorf tube and were preserved in the same manner as fish and
water samples.
DNA extraction and 16S rRNA amplicon sequencing
Gut microbiota from lake sturgeon larvae were surveyed using high-throughput
sequencing of the v4 region of the 16S rRNA gene. The distal gut (spiral valve) of each
sturgeon larvae was recovered from fish following aseptic techniques. The distal gut was
defined as the section extending from the beginning of the intestine until the spiral valve.

22

Exterior surfaces were swabbed with 100% ethanol before dissections of the whole
digestive tract using sterile instruments. Dissections were performed as previously
described by (Milligan-Myhre et al., 2011) with slight modification. The intact GI tracts
were cut from the fish body cavities, and the excised GI tract was immediately
transferred into filtered-sterilized 80% ethanol for DNA isolation. Due to the small size
of the gut, a composite of at least four GI samples of larvae were combined for each tank
replicate, within each treatment group, at each time point. Each tube containing GI tract
samples was first centrifuged for 15min at 4oC to pellet tissues and bacteria before DNA
was extracted. The MoBio PowerSoil® DNA Isolation Kit (Carlsbad, CA, USA)
including a bead-beating step was used following protocols for low-biomass samples as
suggested by the manufacturer with slight modification. The integrity of each DNA
sample was assessed based on the amplification of an approximately 1.4k bp of 16S
rRNA gene (27F and 1389R) followed by gel agarose electrophoresis and DNA
concentration

were

quantified

using

Microplate

spectrophotometer

(BioTek®,

Winooski,VT, USA).
One hundred fifteen DNA samples that have been validated to contain sufficient
bacterial DNA (as shown by the presence of 16S rRNA amplicon bands during
electrophoresis) were submitted for the sequencing at Michigan State University
Research Technology Support Facility, RTSF (East Lansing, MI, USA). All of the
sequencing procedures, including the construction of Illumina sequencing library,
emulsion PCR, and MiSeq paired-end sequencing v2 platforms of the V4 region
(~250bp; primer 515F and 806R) followed standard Illumina (San Diego, CA, USA)
protocols. Michigan State’s Genomics RTSF (https://rtsf.natsci.msu.edu/genomics/)

23

provided standard Illumina quality control; including base calling by Illumina Real Time
Analysis v1.18.61, demultiplexing, adaptor and barcode removal, and RTA conversion to
FastQ format by Illumina Bcl2Fastq v1.8.4.
Sequence processing
Sequence data were processed using default sequencing data analyses pipeline and
computing workflow. Briefly, paired-end sequence merging, quality filtering,
“denoising”, singleton-sequence removal, chimera checking, taxonomic assignments and
Operational Taxonomic Unit (OTU) selection was conducted using an open-source
workflow based on methods implemented by mothur v.1.36.1 (Kozich, Westcott, Baxter,
Highlander, & Schloss, 2013). A reference-based OTU clustering and the taxonomic
assignment was performed using SILVA-based bacterial reference database file provided
in mothur to cluster sequences defined with 97% identity. To minimize effect of under
sampling while maintaining as broad a dataset as possible, the final OTUs table was
rarefied to a depth of 5775 sequences per samples. Five DNA samples with low sequence
depth were discarded in downstream analyses. Rarefaction analyses were performed to
evaluate the sampling coverage for each samples based on the selected sequence depth.
Analyses of bacterial community profiles and ecological statistics analyses
1.1 Alpha diversity
All measures of community diversity and similarity including Inverse Simpson (1/D)
diversity indices and OTUs richness of each sample were calculated from the sequence
data within mothur to quantify alpha [α] diversity. To test for significant differences in
diversity indices among treatment groups (water and diet), and among time periods, a
multiple factor ANOVA was performed on the summary files provided by mothur using

24

the programming and statistical software, R (version 0.98.978) base package. The test
was followed by Tukey HSD post-hoc tests. P-values below 0.05 indicated significant
differences in pairwise mean comparisons.
1.2 Temporal and Differential abundance of OTUs
A custom R code was used to calculate the relative abundance and identify dominant
phyla and taxa (OTUs) in all communities across sampling times (gut microbiota, water,
and Artemia-associated epibiota). Codes were written and implemented using packages
dplyr and reshape2. The relative abundance of all taxa within community samples was
calculated, and taxa with (normalized) abundance exceed 0.1% were considered as ‘predominant’ taxa. The most abundant 20 taxa were subset from total number of OTUs
based on the 0.1% cut-off, and relative abundances were tabled to show temporal
variability in fish gut community composition. The remaining taxa were grouped as
‘Others’.
1.3 Beta diversity
We used several packages implemented in R to perform comparative community
compositional analysis of beta [β] diversity and other community ecological statistics
using the tabulated OTU dataset of pre-dominant taxa. Briefly, we used the vegan
function to generate estimates of Bray-Curtis (BC) distances among sample microbial
communities (Oksanen, 2015). Subsequently, we used the cmdscale function to perform
ordination (Principle Coordinate analyses, PCoA) based on BC distance (Bray & Curtis,
1957). The ggplot and ggplots2 packages (Wickham, 2009) were used to create
ordination plots to visually compare gut and environmental bacterial community
composition among samples collected from different treatments and among sampling

25

periods based on the two largest eigenvalues. Two multivariate hypothesis tests were
implemented using two functions. The adonis function was used to perform multivariate
hypothesis testing on differences between locations of the centroids of treatment group
coordinate ordinations based on Permutational Multivariate Analyses of Variance
(PERMANOVA). The betadisper function was used to perform a Homogeneity of
Multivariate Dispersion (PERMDISP) test on community BC matrices (Anderson, 2001;
2006). These tests were employed because of the non-parametric and skewed nature of
microbial community compositional data. OTUs that had the highest correlation with the
PCoA x and y component axes were identified based on Pearson correlation coefficients
using the corr function.
1.4 Influence of water and diet treatments of GI tract microbiota
To analyze treatment effects of water type and diets on fish gut microbiota, PCoA were
performed separately on fish gut communities for all four-treatment groups within each
developmental stage. Important PCoA axes denoted by comparatively larger eigenvalues
to average eigenvalues were selected. Linear regression models were fit, where each axis
was a response variable given predictor variables of water type and food type. Under the
null hypothesis, we expected gut community composition at each time point to be
unaffected by treatment. Axes represent linear correlations of bacterial taxonomic
composition present in gut communities. Axes that showed significant effects of
treatments or interactions between water and diet treatments were analyzed using the
lsmean function to determine the effects of each factor (or combination of factor(s) in the
interaction) on the bacterial taxonomic composition and relative abundance in larval fish
GI tracts.

26

1.5 Tests evaluating whether GI tract communities were a neutral subset of the
environmental (source) communities
We used a neutral community ecological model adapted from Sloan et al. (2006) to
explore the relative importance of neutral processes (i.e., dispersal and ecological drift)
and selection in gut microbiomes at a given time of sampling (pre-feeding, one week, and
two week active feeding). This model also distinguished members of the gut microbiome
whose presence was consistent with dispersal from surrounding environmental
communities (water as a source) and those that deviated from the neutral model (i.e.,
were over or under represented in the gut relative to the water source(s)). Using
customized R scripts following Venkataraman et al. (2015), we evaluated whether neutral
processes associated for example with probabilities of transport from source communities
were sufficient to explain the observed composition of the gut microbiomes at each
developmental stage. The level of fit of the neutral model served as an adequate
representation of gut community composition if the R2 value was high, whereas the
significance of neutral processes was indicated by the p-value.

27

RESULTS
Sequencing and sample summary
Rarefaction analyses showed that sequencing efforts were consistent across replicate
samples and treatments at a depth of 5775 sequences per sample as denoted by total
percentage coverage higher than 98%. We were able to sample a large portion of the
OTUs and diversity present while still retaining a large number of samples within fish of
a given age. After quality filtering, our 16S rRNA amplicon dataset produced 6,034,269
high quality reads. In total, we observed 4137 OTUs (2894 when omitting singleton
OTUs) defined at 97% sequence identity. From 118 samples submitted for sequencing,
rarefication at 5775 sequences per sample eliminated five samples below this sequencing
coverage.
Morphometric data consisted mean fish weight (g ± SD) and length (mm ± SD)
showing that fish growth in all four treatments was consistent throughout the duration of
the study (data not shown). By the end of the trial, fish raised in stream water grew
significantly larger compared to fish in UV-treated groundwater. No fish health issues
were detected. Survival of fish was nearly 100% in all treatments and developmental
stages.
Ontogenetic changes of dominant bacterial taxa in lake sturgeon larvae rearing
We quantified the number of sequences that were represented by each phylum from GI
tracts of all fish sampled in each developmental period. We found that gut bacterial
communities were comprised of 26 microbial phyla, however the most abundant phyla
covering more than 95% of all sequences, in order of abundance included Proteobacteria,
unclassified phylua, Firmicutes, Bacteroidetes, Actinobacteria, Acidobacteria,

28

Planctomycetes, Cholroflexi, and Fusobacteria. General patterns of bacterial phylumlevel contributions to gut microbiomes were shown with regards to treatment group and
corresponding sampling time (Fig. 1.1). Across all treatment groups, the microbiome
composition shifted from communities dominated by Proteobacteria and several other
phyla prior to initiation of exogenous feeding to Firmicutes dominated communities. The
relative abundance of all other phyla were reduced after fish began actively feeding.
At the genus level pre-feeding gut communities across all treatment groups were
dominated by Proteobacteria from the genus Pseudomonas, comprising between 13 to
23% of total community composition. This genus was also present in food and water,
although at considerably lower levels (~1%). Upon initiation of feeding, several
Firmicutes genera including Sarcina, Exiguobacterium, and Clostridium emerged as the
predominant taxa in the lake sturgeon larval gut microbiota sampled after one week of
active feeding. As fish aged, Sarcina persisted among predominate taxa comprising ~2030% of the gut community in all GW, GWp, S, and Sp groups. However, fish raised in
stream water (S and Sp) had Clostridium as the most predominant taxa after two weeks of
active feeding, whereas Deefgea in the phylum Proteobacteria was as prevalent as
Sarcina in fish raised in the ground water treatment.
We tested whether changes in larval gut microbiota occurred when fish were
raised on experimental diets and rearing water sources across life stages. Over the course
of three weeks of development, we found that microbiome diversity varied among age
cohorts. Overall, community diversity decreased as fish transitioned to active feeding
(Fig. 1.2a and 1.2b). Multiple factor ANOVA analyses quantified sources of variability in
alpha diversity indices (Inverse Simpson diversity), indicated statistically significant

29

differences in diversity as a function of sampling time (F-valueinv=11.31 df=2, p < 0.001),
whereas OTU richness did not differ across sampling times (F-valuerich = 0.91, df=2, p =
0.407) or treatment (F value=1.92, df= 3, p=0.137). No significant interaction was found
between water and food treatments for Inverse Simpson Diversity or richness (F-valueinv
= 1.13, df=6, p=0.358; F-valuerich = 0.38, df=6, p=0.892, respectively).
Variation in gut bacterial community profiles in association with water, diet, and time
To visualize relationships between gut microbial community composition in the gut and
the composition of environmental sources (water and diet), a PCoA was performed to
analyze samples in reduced dimensional space using ordination plots. Variation in
community membership among environmental microbiota and gut microbiota within
each replicates from all treatment groups during each time period indicated agedependent changes in prevalent microbial taxa (Fig. 1.3). Statistical analyses of beta
diversity across sampling periods (fish ages) revealed significant taxonomic
compositional divergence of the microbial community (PERMANOVA test pseudo-F =
4.88, R2 = 0.084, p < 0.001). We reject the null hypothesis of no differences in
multivariate centroid location and microbial community or composition across sampling
times. In addition, the non-significant results of the PERMDISP (p = 0.201) indicated
that compositional dispersion within group was homogenous. Therefore PERMANOVA
results can be interpreted as true differences in time or age-dependent community
composition.
Community composition data were further analyzed to decouple the treatment
effects of water source and diet from the pervasive effects of sampling time. Under the
null hypotheses, water type and food administered treatments were not expected to

30

significantly affect gut community taxonomic composition, and there would be no
interaction between water and diet treatments within each time point. During the prefeeding period (sampling period 1), we found significant water treatment effects on the
PCoA axis 6 since fish has not officially started feeding yet (Table 1.1). As fish
transitioned to active feeding, both water and food treatments interacted significantly to
affect gut community composition after one week of feeding (PCoA axis 1, Fig. 1.4a).
The effect of water treatment was significant on the PCoA axis 5 (Table 1.1). During
sampling period 3 after the second week of active feeding, gut community composition
was statistically different between water treatments. There was no effect of diet (PCoA
axis 1, Fig. 1.4b). Least-Square means (LS-mean) values for all-important axes across all
time are shown in Table 1.1.
Focusing on the first coordinate axis during sampling periods 2 and 3 following
initiation of feeding, we investigated whether there was evidence for co-occurrence of
taxa in gut communities. Genus Clostridium (Otu001) associated with family
Clostridiaceae (phylum Firmicutes) was found to have strong, positive correlation with
the first axis during the first week of feeding (Pearson correlation, r = 0.884). During the
second week of feeding, two other bacterial taxa (Sarcina,(Otu002) and Unclassified
genera (Otu004), both members of Clostridiaceae family) showed a strong correlation
(Sarcina r = -0.791, Unclassified Clostridiales r = 0.743) with the first coordinate axis.
Another taxon identified as genus Deefgea, a member of the family Neisseriaceae
(phylum Proteobacteria, order Gammaproteobacteria) was also in high abundance and
was positively correlated on the same axis (r = 0.779).

31

Neutral processes are not the dominant mechanisms generating and maintaining
community composition during early gut microbiome assembly
LS-mean analyses of PCoA axis 1 at sampling point 3 (two weeks after initiation of
feeding, 22 dph) suggest that dispersal of bacteria from environmental (water) sources
into the gut was important in shaping the community composition in the gut. However,
previous analyses did not address the question Is the fish gut community at each
developmental stage a neutral subset of the source community? We wished to distinguish
between species that are detected in the gut because of neutral processes and
deterministic processes that may be associated with age-dependent gut conditions. To
answer this question, we applied the neutral model theory based on the Sloan (2006)
Neutral Model for Prokaryotes to investigate developmental processes underlying gut
microbiome compositional and successional change over time. The neutral model
assumes that community composition can be explained by the dispersal of microbial taxa
from the surroundings and ecological drift (stochastic change) within the source
community (Venkataraman et al. 2016). Based on LS-mean analyses, we found that water
was a more important source of gut community members than was diet.
Neutral processes were detected as part of important contributors to microbial
community formation during each sampling period (developmental stage) as indicated by
significant p-values (Figs. 1.5a-c). These processes are significant, yet not strongly
predictive. Using R2 as a measure of goodness-of-fit, we determined that gut community
composition from across all treatments poorly fit a neutral model based on the low R2
value of model fit.

32

Across sequential sampling periods, as fish aged, the number of shared microbial
OTUs between the gut and water decreased (Table 1.2). At the pre-feeding stage, larval
GI tracts contained the most neutrally dispersed taxa (114), but this number declined
dramatically as fish began active feeding (45). Two weeks after active feeding, only 38
taxa were shared between gut and water communities, and about half of these taxa were
under-represented or over-represented in the gut (Table 1.2). We conclude from these
analyses that the OTUs of gut microbiomes are likely under selection. Taxa present in the
gut are not a completely neutral subset of the taxa present in the source water
communities.

33

DISCUSSION
Many basic microbial community ecology questions, concerning the dynamics of
community composition at the onset of colonization and early ontogenetic changes, are of
importance to understand host-microbial relationships in the wild and under domestic
conditions. Information pertaining to the source of microbial communities that establish
in the fish gut, neutral vs non-neutral dynamics during community assembly and generate
inter-individual variability, and the interaction of animal host growth with the
establishment of gut communities that collectively affect taxonomic diversity and relative
abundance are incompletely known.
In this study, we were interested in characterizing the development of fish gut
microbiomes during important early developmental stages. Our experimental system
enhanced understanding of the factors affecting initial colonization and development of
the lake sturgeon larval gut microbiota prior to and during the critical transition from
endogenous to exogenous feeding. Our research contrasts previous studies on other
sturgeon species associated with diet and gut microbiota that were either performed on
larger/older fish. For example,e.g., Geraylou et al. (2013) and Bacanu & Opera (2011)
used from Siberian sturgeon (Acipenser barii) and white sterlet sturgeon (Acipenser
ruthenus), respectively at the juvenile stage ( average weight between 15g-30g and more
than 3 months post hatch). Other researchers (e.g., Askarian et al., 2011; Bacanu &
Oprea, 2013; Callman & Macy, 1984; Geraylou et al., 2013; Masouleh et al., 2006) used
fish that had been exposed to cultivation for extended periods.
To our knowledge, this report provides the first experimental system to evaluate
environmental (focusing both on rearing water type and diet administered) and host-

34

associated factors that affect the compositional dynamics of the intestinal microbiota
using high-throughput sequencing based on 16S rRNA specifically in sturgeon, an
important aquaculture teleost fish species and a species of conservation concern. These
factors have a profound influence on microbial membership and could further impact the
metabolic potential of the gut microbiome. Advancing uses of microbial manipulation
(pre- and probiotics) is a goal for the aquaculture industry to promote fish growth and
health. Accordingly, advancing understandings of compositional dynamics that naturally
occur in the gut microbiota of cultured fish species like sturgeon has relevance to
commercial and conservation aquaculture. Here, we provide evidence that changes in
microbiota in the gut of lake sturgeon occurs early in development. Gut microbiota
composition quickly diverged from the environmental communities associated with their
respective rearing water and diet conditions.
Lake sturgeon larvae were raised in constant, controlled environments through
manipulation of water and diet over three sampling periods from 11 to 29 dph. Our
results suggest that major compositional shifts in gut community composition
corresponding to different developmental stages. Taxonomic profiles changed across
sequential larval stages extending before the onset of exogenous feeding (pre-feeding)
through 29 dph when fish were actively feeding, and when GI tract structures resembled
adult anatomical structures (Buddington & Christoferson, 1985; Buddington &
Doroshov, 1986). Findings suggest that during these early ontogenetic periods, host
physiological development likely serves as a strong deterministic force directing the
formation of gut communities, regardless of food type or surrounding (water)
environmental communities. The temporal shift in bacterial community composition has

35

also been documented in others studies involving zebrafish and rainbow trout
Onchorhychus mykiss) larvae during periods of constant diet and environmental
condition. However, while rainbow trout studies reported the strong influence of diet type
and environmental factors to gut microbiota composition, findings from the zebrafish
studies differed (Stephens et al., 2015, Wong et al. 2015, Ingerslev et al., 2014).
Pronounced temporal changes in microbial community composition occurred
from the pre-feeding to exogenous feeding periods. Proteobacteria dominated gut
communities during the pre-feeding stage, whereas Firmicutes dominated communities
after two weeks of feeding. Large decreases in community taxonomic diversity were also
documented between pre- and post-feeding periods. Results are concordant with findings
for other fish species, that documented Proteobacteria and/or Firmicutes as among the
most abundant phyla in fish gut communities. In zebrafish, studies by Stephens et al.
(2015) indicated that Firmicutes and Proteobacteria are most common in larvae, but
adult microbiota were dominated by Fusobacteria (Rawl, 2006). Studies on microbiome
ontogenetic shifts in rainbow trout fry also indicated the presence of Proteobacteria or
Firmicutes associated with either marine-based or freshwater plant-based diet offered
during first feeding (Ingerslev et al., 2014). Geraylou et al. (2013) found that
administration of prebiotic to Siberian sturgeon (Acipenser barii) shifted gut microbiota
primarily in the phylum Firmicutes. Comparative analyses of gut microbiota from 8
freshwater fish species encompassing fish with different feeding habits revealed that
Proteobacteria and Firmicutes were the dominant phyla in all fish species (Li et al.,
2014a). Another cross-sectional gut microbiome study performed by Bledsoe et al.

36

(2016) on channel catfish also indicated the prevalence of Proteobacteria in fish at 3 dph,
and the appearance of Firmicutes along with Proteobacteria when fish reached 65 dph.
We also found that predominant taxnomic assemblages tended to consist of
closely related taxa from the family Clostridiaceae. This includes Sarcina, Clostridium,
and unclassified genera from the family. This might indicate that traits underlying
assemblage membership were often shared among related organisms. One of these taxa,
genus Clostridium is enriched following the transition to active feeding. In humans, this
genus is part of important commensal microbiota that begin to colonize human intestines
of breast fed infants as early as the first month of life, and have been shown to play roles
in modulating gut homeostasis over the entire life span (see review by Lopetuso et al.
2013). Studies in mice indicated that commensal Clostridia populate specific regions in
the intestinal mucosa, thus establishing a close relationship with gut cells and perform
critical physiological functions (Lopetuso et al., 2013).
Another taxa, Sarcina was also found to be abundant by the time fish reached 22
dph and actively feeding. Sarcina are fermenting bacteria that are frequently found in the
gastric contents and faeces of human patients with gastro-intestinal disorders. All the
strains were obligate anaerobes, fermented cellulose and required a carbohydrate for
growth, all produced ethanol but not butyric acid from glucose and can also produced
acid and gas from sugar like glucose, fructose, sucrose, maltose, lactose, galactose and
raffinose (Crowther, 1971). One species, S. ventriculi is widespread in the soil and may
be considered to be part of the intestinal flora of human although its significance remains
unknown (Crowther, 1971; Smit, 1911). Given the biology of both taxa and feeding
nature of sturgeon as a bottom feeder, it is possible to have Sarcina and Clostridium as

37

predominant taxon in their distal gut. Sturgeon possess a valvular hindgut (spiral valve)
that serves as the primary region of digestion and nutrient absorption, and thus might
provide an abundance of nutrients for bacterial like Sarcina to flourish (Buddington &
Christofferson, 1985; Callman & Macy, 1984) and Clostridium to maintain all gut
physiological functions. Studies by Callman & Macy (1984) also showed that anaerobic
bacterial fermentation takes place in the spiral valve producing volatile fatty acids
(VFA’s) and hydrogen gas as by-product, supporting the idea that these bacteria may
enhance digestive efficiency.
Two important Proteobacteria taxa were Pseudomonas and Deefgea.
Pseudomonas, was present during all stages, although it was initially present in
abundance during the pre-feeding stage and later became less abundant at one and two
weeks after active feeding began. The presence of Pseudomonas in relatively high
abundance has also been reported as part of lake-sturgeon egg-associated community
(Fujimoto et al., 2013b). Romero & Navarrete (2006) found Pseudomonas sp. present in
abundant within gastrointestinal tracts of fish juvenile as well as on eggs, but not in the
water nor in food when they studied bacterial communities associated with early life
stages in coho salmon (Onchorynchus kisutch). This is likely due to vertical transmission
of a pioneering strain from eggs to fish GI tracts. Pseudomonas is also commonly
observed in gut microbiota of mature fish (Hansen & Olafsen, 1999; Llewellyn et al.,
2014; Navarrete et al., 2008). A number of diet related studies have reported variability in
relative abundance of Pseudomonas sp. that were affected by differential food treatment.
These taxa are invariably influenced by experimental diet including dietary inulin in
surubins (Mourino et al. 2012), dietary mannanoligosaccharides (MOS) in rainbow trout

38

(Dimitrolou et al. 2009), and dietary yeast culture in both hybrid and Nile tilapia (He et
al. 2010)(see review by Ringo et al. 2016).
Little information is available pertaining to Deefgea sp. Previous studies by Chen
et al. (2010) reported so far only two species found in genus Deefgea (Family
Neisseriaceae, Order Betaproteobacteria). Those species were described as Deefgea
rivuli and Deefgea chitinilytica (Chen et al. 2010; Stackebrandt et al. 2007). However,
both taxa originated from hard-water and wetland samples, respectively. In another study,
Jung & Jung-Schroers (2011) first documented the association of Deefgea isolates with
fish. They reported six isolates of D. chitinilytica were cultured from swabs of skin and
internal organs of two freshwater ornamental fish species farm-raised in farm [gold tench
(Tinca tinca) and goldfish (Carassius auratus auratus)]. Several other bacterial taxa from
the same family were described as chitin-hydrolyzing species, and D.chitinilytica was
suspected to have similar function too. Due to this, Deefgea could have significance
importance listed to be among opportunistic taxa that could play roles in infections of
aquatic organism (Jung & Jung-Schroers, 2011).
Throughout the assembly of early gut bacterial communities, stochastic and
deterministic factors associated with water and food epibiota could play roles in shaping
these ecological communities (Llewellyn et al., 2014; Navarrete et al., 2009; Romero &
Navarrete, 2006). As we have shown in ANOVA and multivariate analyses of gut
microbiome at each time point, the water rearing environment initially had a strong
influence on community composition during the pre-feeding stage, as gut communities
reflected aquatic communities of both water treatments. Water appears to serve as
primary inoculant before and during the transition stage from endogenous to exogenous

39

feeding. Early stages of gut microbiota colonization were temporally unstable and
stochastic processes such as random recruitment of water epibiota into the gut occurred at
the beginning of exposure of the early upper GI tract and gill surface to ambient water
(Hansen & Olafsen, 1999). As fish developed further, gut communities changed
coincident with initiation of active feeding and diverged in composition from to
surrounding communities. Previous studies performed by our lab on egg surface of lake
sturgeon also displayed directional changes of bacterial community along with egg
developmental stages (Fujimoto et al. 2013). Another study by Giatsis and colleagues
(2015) also documented changes in the gut community structure over time with
significant contributions of water bacterial communities.
As feeding continued, significant interactions were observed. We observed no
significant differences between gut communities of supplemented vs non-supplemented
food treatment groups within the stream water environment. In contrast, significant
differences were observed between gut microbiota of fish raised in groundwater based on
the food treatment administered (supplemented vs non-supplemented diet). Gut
communities of fish raised in groundwater differed according to diet, indicating that diet
influences gut community membership. However, further analyses showed that the effect
of diet on gut community composition after the pre-feeding stage was not evident. Only
water treatment was significantly associated with gut community composition at two
weeks of active feeding. At this point, gut communities differed between stream and
ground water treatments.
Dynamics of community compositional change during the gut assembly processes
can be attributed many processes (Costello et al. 2012). First, the composition of the gut

40

community could be determined by environmental selection. In this study, the
environments were associated with either the rearing condition of fish, as fish raised
within similar water environments were exposed to similar pools of microbial taxa
present in each source. Rearing temperatures of surface water and ground water may also
have contributed to gut-selective environments. Sturgeon raised in our hatchery exhibited
faster growth rate in surface water with temperature relatively warmer than ground water,
supporting previous report that rates of sturgeon during early development are
temperature dependent (Wang et al. 1985). The local community could also be under the
influence of neutral vs selective processes. In the first situation, taxa present in the gut
could be a random draw of species present in water. Subsequently, maturation of the fish
gut could impose selective pressures that favor particular subsets of taxa, or inhibit the
growth of certain subsets of taxa (Giatsis et al., 2015).
The evaluation of the relative importance of neutral vs deterministic processes
was achieved through implementation of a neutral model. Our findings suggested that
neutrality (e.g., random dispersal of taxa from water) does occur but is not the pervasive
ecological force shaping the gut community during all three developmental stages. Low
model R2 values indicated that neutral processes are not dominant. Deterministic
processes are also likely shaping changes in gut community composition associated with
development of fish GI tract as evidenced by reductions in the number of taxa between
the gut and water (Figs. 1.5a-c), even though the proportional contributions of over- and
under-represented taxa remained approximately unchanged across periods.
Morphometric data within sampling period revealed that length and weight of fish
larvae differed significantly among water treatment as ambient water temperature

41

affected growth. Fish in groundwater developed more slowly compared to fish reared in
stream water during those ontogenetic stages. As a result, their gut offered different
microenvironment that could support certain microbial taxa, which in turn affected
microbial relative abundance and taxonomic diversity.
We recommend future studies include experimental designs that sample over a
longer duration, including manipulations of the water community. Further studies
focusing on taxa that increase in relative abundance over time would be useful to
establish ecological functions that could be used for greater efficiency in applications of
microbial-based fish management in aquaculture. Such data could also be used to identify
suitable release sites for fish from conservation hatchery programs.
Overall, the significant changes in diversity and taxonomic composition of the
lake sturgeon gut microbiomes occurred principally associated with early developmental
stages in connection with the initiation of first feeding. Microbial communities diverged
following initiation of exogenous feeding compared to a community of surrounding water
and food epibiota. Our understanding and abilities to control (i.e., application of
probiotics) underlying deterministic and stochastic factors associated with the source of
microbial innocula appeared to be in part tied to communities in the water when feeding
begin. Future studies may profitably explore effects of manipulations of communities in
rearing water and food samples to understand the dynamics of microbial community
assembly associated with these factors.

42

APPENDIX

43

APPENDIX
Table 1.1 showing R2-values, p-values, and Least-square mean values of least-square mean analyses performed on significant PCo
axes for each sampling period. Five axes across all three stages that showed significant effect of either water, food, or interaction of
both water and food treatments on the microbial community composition were Axis 6 for pre-feeding, Axis 1 and Axis 5 for stage at
one-week post-active feeding, Axis 1 at two-week post-active feeding. Interaction plots were shown in Figure 1.4
Least-square means
Linear regression model :
Pco Axes ~ Water type * Food administered
Important
axes
Signif. Treatment
p-value
R2 value
Pre-feeding
Pco Ax.6
Water type
0.045
0.206
One wk
active
feeding
Interaction (Water type
Pco Ax. 1
and food type)
0.004
0.444
One wk
active
feeding
Pco Ax.5
Water type
0.016
0.233
Two wk
active
feeding
Pco Ax.1
Water type
0.020
0.267

Stream Water

Ground Water

Food - Suppl.

Food - No suppl.

Food - Suppl.

Food - No suppl.

0.044±0.06

-0.137±0.05

0.070±0.05

0.030±0.05

-0.079±0.09

-0.048±0.09

0.327±0.09

-0.300±0.12

0.064±0.05

0.079±0.05

-0.049±0.05

-0.142±0.06

-0.189±0.10

-0.125±0.10

0.086±0.10

0.227±0.10

44

Table 1.2 Results of neutral models applied to gut microbiota showing the number and proportion of shared OTUs detected in both
thegut microbial communities and potential microbial source (water). Over-represented taxa are those that were selected for (i.e.,
abundance low in water, but detected in higher abundance in the gut) whereas under-represented taxa are present in lower abundance
in the gut than in water.
Developmental stages
Total number of shared OTUs between source
(rearing water) and target (fish gut) communities
Number of neutrally dispersed OTUs
(Proportion, %)
Number of overrepresented OTUs
(Proportion, %)
Number of underrepresented OTUs
(Proportion, %)

Pre-feeding
160
114
(71.25%)
31
(19.38%)
15
(9.37%)

45

One-week after
active feeding
45
27
(60.00%)
11
(24.44%)
7
(15.56%)

Two-week after
active feeding
38
22
(57.89%)
12
(31.58%)
4
(10.53%)

(a)

(b)

Figure 1.1 Bacterial composition of different communities identified from (a) lake
46

Figure 1.1 (cont’d)
sturgeon larval gut and (b) environmental samples. (a) Relative abundance of
dominant bacterial phyla found in lake sturgeon larval gut microbiota across treatment
and during different developmental stages. Only the dominant phyla were shown in the
bar chart (Acidobacteria, Actinobacteria, Bacteroidetes, Firmicutes, Fusobacteria,
Proteobacteria) and the rest of taxa were assigned as Others. All four treatment groups
were denoted with acronyms S, Sp, GW, GWp (b) Relative abundance of dominant
bacterial phyla found in environmental microbiota. Only the dominant phyla were
shown in the bar chart (Acidobacteria, Actinobacteria, Bacteroidetes, Firmicutes,
Proteobacteria, Verrucomicrobia) and the rest of taxa were assigned as Others.
Classification for environmental communities were denoted with acronym Artemia-GW,
Artemia-S, Detritus, Water-GW, Water-S
S (fish raised in natal stream, fed live Artemia);
Sp (fish raised in natal stream, fed live Artemia mixed with retentate);
GW (fish raised in groundwater, fed live Artemia);
GWp (fish raised in groundwater, fed live Artemia mixed with retentate)
Artemia-GW. (food treatment in which live Artemia was prepared using ground water);
Artemia-S (food treatment in which live Artemia was prepared using stream water);
Detritus. (food treatment in which retentates were collected from sock-filter prior to mixing with live Artemia);
Water-GW (water treatment - groundwater that has been UV-filtered);
Water-S (water treatment – stream water)

47

(a)

(b)

Figure 1.2 Estimates of Alpha diversity for lake sturgeon gut microbial communities
48

Figure 1.2 (cont’d)
from all treatments across all developmental stages. Statistics comparing gut
microbiotas across treatment at different ages were made using a two-way ANOVA.
Points indicate mean value of diversity index, colored by different treatments. All four
treatment groups were denoted with acronym S, Sp, GW, GWp
(a) Alpha diversity in gut microbiota at each time point, as measured by Inverse Simpson
indice.
(b) OTU richness based on number of taxa observed in gut microbiota from all treatment
and times.
S (fish raised in natal stream, fed live Artemia);
Sp (fish raised in natal stream, fed live Artemia mixed with retentate);
GW (fish raised in groundwater, fed live Artemia);
GWp (fish raised in groundwater, fed live Artemia mixed with retentate)

49

0.4

PCoA2, Var.. explained = 7.2%.

Time
Pre feeding
1 wk active feeding
2 wk active feeding

0.2

Communities
GW
GWp
S

0.0

Sp
Environment

0.2

0.2

0.0

0.2

0.4

PCoA1, Var. explained = 21.5%
Figure 1.3 Visual representation of ontogenetic changes in lake sturgeon larval gut
microbiota based on Bray-Curtis distances. Bray-Curtis dissimilarities between
communities originating from gut, water and food samples from all four different
treatments at three developmental stages (pre-feeding, a week active feeding, two
week active feeding) were visualized using Principal coordinates analyses (PCoA, or
also known as Metric Dimensional Scaling MDS) plots. Points represent the samples
from each replicate. Fish gut microbial community from four treatment groups were
denoted by acronyms S, Sp, GW, GWp, whilst microbial communities from water and
food collectively denoted by Environment
S (fish raised in natal stream, fed live Artemia);
Sp (fish raised in natal stream, fed live Artemia mixed with retentate);
GW (fish raised in groundwater, fed live Artemia);
GWp (fish raised in groundwater, fed live Artemia mixed with retentate)

50

(a)

(b)

Figure 1.4 Interaction plots of marginal (least-square, LS) means for first PCoA
axes (axes that explained largest variation in dataset) at different developmental
stages.

51

Figure 1.4 (cont’d)
Additional information pertaining to the LS means was compiled in Table 1.1. (a) First
PCo Axis at one week active feeding; (b) First PCo Axis at two week active feeding.
Figure 1.4(a) indicates significant interactions occurred between food and water
treatments. Water and food treatments has influenced on the gut microbial community
composition (represented by first PCo axis) during the sampling period at the first week
post-active feeding. No significant difference in gut community composition between
supplemented vs non-supplemented food treatments within the stream water
environment. However, significant difference was observed between gut microbiota of
fish raised in groundwater based on food treatment (Fig. 1.4(b)). Significant effects of
water treatment on gut composition (represented by first PCo axis) were observed in fish
at two weeks after active feeding began. Diet effects were no longer observed.
Food – suppl. (food treatment in which live Artemia mixed with retentate was offered to the fish);
Food – no. suppl. (food treatment in which live Artemia only was offered to the fish);

52

(a)

(b)

(c)

Figure 1.5 Results of neutral model testing with water as the source of gut microbial communities (a) at pre-feeding stage, (b)
after one-week active feeding, (c) after two weeks of active feeding. The solid black line represents the best-fit neutral model
generated using a beta probability distribution. The model is developed based only on taxa described in both gut and water sources.
The dashed lines represents 95% confident intervals around the best-fitting neutral model. Species within the confidence intervals
(gray points) are classified as neutrally dispersed taxa, likely present in the gut as a result of neutral processes (such as dispersal, or
ecological drift). Species deviating from neutral model (red and green points) were classified as under-represented or over-represented
taxa, respectively. These taxa are likely affected by deterministic processes such as selection, or could have differential dispersal
ability compared to other taxa in the water. The coefficient of determination (R2) represents the goodness of fit of relative abundance
under the neutral model. The value ranges from ≤ 0 (no fit) to 1 (perfect fit). The P-value indicates that neutral processes that were
detected are significant and not by random chance. In general, neutrality could not be rejected during all three developmental stages,
but fit of data to neutrality expectations was poor as shown by the relatively low R2 values.

53

LITERATURE CITED

54

LITERATURE CITED
Anderson, M. J. (2001a). A new method for non-parametric multivariate analysis of
variance, Australian Ecology, 26: 32-46
Anderson, M. J. (2006). Distance-based tests for homogeneity of multivariate
dispersions. Biometrics, 62: 245-253
Anderson, M. J. and Walsh, D. C. I. (2013). What null hypothesis are you testing?
PERMANOVA, ANOSIM, and the Mantel test in the face of heterogenous
dispersions. Ecology monographs, 83: 557-574
Askarian, F., Kousha, A., Salma, W., & Ringo, E. (2011). The effect of lactic acid
bacteria administration on growth, digestive enzyme activity and gut microbiota in
Persian sturgeon (Acipenser persicus) and beluga (Huso huso) fry. Aquaculture
Nutrition, 17, 488–497.
Askarian, F., Kousha, a., & Ringø, E. (2009). Isolation of lactic acid bacteria from the
gastrointestinal tract of beluga ( Huso huso ) and Persian sturgeon ( Acipenser
persicus ). Journal of Applied Ichthyology, 25, 91–94.
Bacanu, G. M., & Oprea, L. (2013). Differences in the gut microbiota between wild and
domestic Acipenser ruthenus evaluated by Denaturing Gradient Gel Electrophoresis.
Romanian Biotechnology Letters, 18(2): 8069–8076.
Bauman, J.M. (2015). Investigations of aquaculture methodologies to enhance success of
Great Lakes Lake Sturgeon Streamside Facilities. Master Thesis, Department of
Fisheries & Wildlife, Michigan State University, East Lansing, Mich., USA.
Bledsoe, J. W., Peterson, B. C., Swanson, K. S., & Small, B. C. (2016). Ontogenetic
Characterization of the Intestinal Microbiota of Channel Catfish through 16S rRNA
Gene Sequencing Reveals Insights on Temporal Shifts and the Influence of
Environmental Microbes. Plos One, 11(11): e0166379.
Bray, J. R. and Curtis, J. T. (1957) An ordination of the upland forest communities of
southern Wisconsin. Ecological monographs. 27(4): 325-49
Buddington, R. K., & Christofferson, J. P. (1985). Digestive and feeding characteristics
of the chondrosteans. Environmental Biology of Fishes, 14(1): 31–41.
Cain, K., & Swan, C. (2010). Barrier function and immunology. In Grossell, M., Farrel,
A. P., Brauner, J. (Eds.) The Multifunctional Gut of Fish : Fish Physiology. Vol. 30,
pp: 111-134. Academic Press, Elsevier, London

55

Callman, J. L., & Macy, J. M. (1984). The predominant anaerobe from the spiral intestine
of hatchery-raised sturgeon (Acipenser transmontanus), a new Bacteroides species.
Archive of Microbiology, 140: 57–65.
Chen, W.M., Chung, Y.N., Chiu, T.F., Cheng, C.Y., Arun, A.B. and Sheu, S.Y. (2010)
Deefgea chitinilytica sp. nov., isolated from a wetland. International Journal of
Systematic and Evolutionary Microbiology, 60: 1450–1453.
Christian, N., Whitaker, B. K., & Clay, K. (2015). Microbiomes: unifying animal and
plant systems through the lens of community ecology theory. Frontiers in
Microbiology, 6(869): 1-19
Costello, E. K., Stagaman, K., Dethlefsen, L., Bohannan, B. J. M., & Relman, D. A.
(2012). The Application of Ecological Theory Toward an Understanding of the
Human Microbiome. Science, 336(6086): 1255–1262.
Crowther, J. S. (1971). Sarcina ventriculi in human feces. Journal of Medical
Microbiology, 4: 323-350
Dave, M., Higgins, P. D., Middha, S., & Rioux, K. P. (2012). The human gut
microbiome: Current knowledge, challenges, and future directions. Translational
Research, 160(4): 246–257.
De Schryver, P., & Vadstein, O. (2014). Ecological theory as a foundation to control
pathogenic invasion in aquaculture. The ISME Journal, 1–9.
Deng, D.-F., Koshio, S., Yokoyama, S., Bai, S. C., Shao, Q., Cui, Y., & Hung, S. S. .
(2003). Effects of feeding rate on growth performance of white sturgeon (Acipenser
transmontanus) larvae. Aquaculture, 217(1–4): 589–598.
Dettlaff, T. A., Ginsburg, A. S. & Shmalhausen, O. I. (1993). Sturgeon Fishes
Developmental Biology and Aquaculture. Berlin.
Diggle, P.K. 1994. The expression of andromonoecy in Solanum hirtum (Solanaceae):
phenotypic plasticity and ontogenetic contingency. American Journal of Botany
81(10):1354-1365.
Dimitroglou A., Merrifield D.L., Moate R., Davies S.J., Spring P.,Sweetman J. &
Bradley G. (2009) Dietary mannan oligosaccharide supplementation modulates
intestinal microbial ecology and improves gut morphology of rainbow trout,
Oncorhynchus mykiss (Walbaum). Journal of Animal Science, 87: 3226–3234.
Fujimoto, M., Crossman, J. A., Scribner, K. T., & Marsh, T. L. (2013). Microbial
Community Assembly and Succession on Lake Sturgeon Egg Surfaces as a Function
of Simulated Spawning Stream Flow Rate. Microbial Ecology, 66(3): 500–511.

56

Gensollen, T. Iyer, S. S., Kasper, D. L., & Blumberg, R. S. (2016). How colonization by
microbiota in early life shapes the immune system. Science. 352(6285): 539-544
Geraylou, Z., Souffreau, C., Rurangwa, E., Maes, G. E., Spanier, K. I., Courtin, C. M., …
Ollevier, F. (2013). Prebiotic effects of arabinoxylan oligosaccharides on juvenile
Siberian sturgeon (Acipenser baerii) with emphasis on the modulation of the gut
microbiota using 454 pyrosequencing. FEMS Microbiology Ecology, 86(2): 357–
371.
Ghanbari, M., Kneifel, W., & Domig, K. J. (2015). A new view of the fish gut
microbiome: Advances from next-generation sequencing. Aquaculture, 448: 464–
475.
Giatsis, C., Sipkema, D., Smidt, H., Verreth, J., & Verdegem, M. (2014). The
Colonization Dynamics of the Gut Microbiota in Tilapia Larvae. PLoS ONE, 9(7),
e103641.
Giatsis, C., Sipkema, D., Smidt, H., Verreth, J., & Verdegem, M. (2015). The impact of
rearing environment on the development of gut microbiota in tilapia larvae.
Scientific Reports, 5(18206): 1–15.
Givens, C. E., Ransom, B., Bano, N., & Hollibaugh, J. T. (2015). Comparison of the gut
microbiomes of 12 bony fish and 3 shark species. Marine Ecology Progress Series,
518: 209–223.
Gómez, G. D., & Balcázar, J. L. (2008). A review on the interactions between gut
microbiota and innate immunity of fish. FEMS Immunology and Medical
Microbiology, 52(2): 145–154.
Hansen, G. H., & Olafsen, J. A. (1999). Bacterial Interactions in Early Life Stages of
Marine Cold Water Fish. Microbial Ecology, 38(1): 1–26.
Hattori, M., & Taylor, T. D. (2009). The human intestinal microbiome: a new frontier of
human biology. DNA Research : An International Journal for Rapid Publication of
Reports on Genes and Genomes, 16(1), 1–12.
He, S., Zhou, Z., Meng, K., Zhao, H., Yao, B., Ringø, E. & Yoon, I. (2010). Effects of
dietary antibiotic growth promoter and Saccharomyces cerevisiae fermentation
product on produc- tion, intestinal bacterial community, and nonspecific immunity
of hybrid tilapia (Oreochromis niloticus female x Oreochromis aureus male).
Journal of Animal Science, 89: 84–92.
Hooper, L. V., Littman, D. R., & Macpherson, A. J. (2012). Interactions Between the
Microbiota and the Immune System. Science, 336(6086): 1268–1273.
Hubbell SP. (2001). The unified neutral theory of biodiversity and biogeography,

57

p 29. Princeton University Press, Princeton, NJ.
Huey, R. B., Kearney, M. R., Krockenberger, A., Holtum, J. A. M., Jess, M., & Williams,
S.E. (2012). Predicting organismal vulnerability to climate warming: roles of
behaviour, physiology, and adaptation. Philosophical Transactions of the Royal
Society Biological Sciences, 367:1665-1679.
Ingerslev, H-C., Strube, M. L., von Gersdorff Jørgensen, L., Dalsgaard, I., Boye, M., &
Madsen, L. (2014). Diet type dictates the gut microbiota and the immune response
against Yersinia ruckeri in rainbow trout (Oncorhynchus mykiss). Fish & Shellfish
Immunology, 40: 624-633
Jung, A. & Jung-Schroers, V. (2011). Detection of Deefgea chitinilytica in freshwater
ornamental fish. Letters in Applied Microbiology, 52(5): 497-500
Kostic, A. D., Howitt, M. R., & Garrett, W. S. (2013). Exploring host – microbiota
interactions in animal models and humans. Genes & Development, 27: 701–718.
Kozich, J. J., Westcott, S. L., Baxter, N. T., Highlander, S. K., & Schloss, P. D. (2013).
Development of a dual-index sequencing strategy and curation pipeline for
analyzing amplicon sequence data on the miseq illumina sequencing platform.
Applied and Environmental Microbiology, 79(17): 5112–5120.
Lederberg, B. J., & McCray, A. T. (2001). ’ Ome Sweet ’ Omics-- A Genealogical
Treasury of Words. The Scientist, 15(7), 8.
Li, J., Ni, J., Li, J., Wang, C., Li, X., Wu, S., Zhang, T., Yu, Y., & Yan, Q. (2014).
Comparative study on gastrointestinal microbiota of eight fish species with different
feeding habits. Journal of Applied Microbiology, 117(6): 1750–1760.
Llewellyn, M. S., Boutin, S., Hoseinifar, S. H., & Derome, N. (2014). Teleost
microbiomes: The state of the art in their characterization, manipulation and
importance in aquaculture and fisheries. Frontiers in Microbiology, 5 (207):1-17.
Lopetuso, L. R., Scaldaferri, F., Petito, V., & Gasbarrini, A. (2013). Commensal
Clostridia: leading players in the maintenance of gut homeostasis. Gut Pathogens,
5(23): 1-8
Macpherson, A. J., & Harris, N. L. (2004). Interactions between commensal intestinal
bacteria and the immune system. Nature Reviews. Immunology, 4(6): 478–485.
Masouleh, A. S., Sharifpour, I., & Arani, A. S. (2006). The aerobic bacterial flora of
hatchery reared Caspian Sea sturgeon fingerlings. Journal of Applied Ichthyology,
22(S1): 261–264.

58

Milligan-Myhre, K., Charette, J. R., Phennicie, R. T., Stephens, W. Z., Rawls, J. F.,
Guillemin, K., & Kim, C. H. (2011). Study of host-microbe interactions in zebrafish.
In Methods in cell biology, 105: 87-116
Mim, S. D., Lazur, A., Shelton, W. L., Gomelsky, B., & Chapman, F. (2002). Production
of sturgeon-South Regional Aquaculture Center.pdf. Southern Regional Aquaculture
Center, 7200: 1-8
Navarrete, P., Espejo, R. T., & Romero, J. (2009). Molecular analysis of microbiota
along the digestive tract of juvenile atlantic salmon (Salmo salar L.). Microbial
Ecology, 57(3): 550–561.
Navarrete, P., Mardones, P., Opazo, R., Espejo, R., & Romero, J. (2008). Oxytetracycline
treatment reduces bacterial diversity of intestinal microbiota of Atlantic salmon.
Journal of Aquatic Animal Health, 20(3): 177–183
Nayak, S. K. (2010). Role of gastrointestinal microbiota in fish. Aquaculture Research,
41(11): 1553–1573.
Nemergut, D. R., Schmidt, S. K., Fukami, T., O'Neill, S. P., Bilinski, T. M., Stanish, L.
F., Knelman, J. E., Darcy, J. L., Lynch, R. C., Wickey, P. and Ferrenberg, S. (2013).
Patterns and Processes of Microbial Community Assembly. Microbiology and
Molecular Biology Reviews, 77(3):342-356
Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O'Hara, R. B.,
Simpson, G. L., Solymos, P., Stevens, M. H. H., and Wagner, H. (2015). vegan:
Community Ecology Package. R package version 2.2-1.
Orizaola, G., E. Dahl, and A. Laurila. 2010. Compensating for delayed hatching across
consecutive life-history stages in an amphibian. Oikos, 119: 980-987.
Rawls, J. F., Samuel, B. S., & Gordon, J. I. (2004). Gnotobiotic zebrafish reveal
evolutionarily conserved responses to the gut microbiota. Proceedings of the
National Academy of Sciences of the United States of America, 101(13): 4596–601.
Rawls, J. F., Mahowald, M. A., Ley, R. E., & Gordon, J. I. (2006). Reciprocal gut
microbiota transplants from zebrafish and mice to germ-free recipients reveal host
habitat selection. Cell, 127(2): 423-433.
Ringø, E., Zhou, Z., Vecino, J. L. G., Wadsworth, S., Romero, J., Krogdahl, J., Olsen, R.
E., Dimitroglou, A., Foey, A., Davies, S., Owen, M., Lauzon, H. L., Martinsen, L.
L., Bossier, P., Sperstad, S., and Merrifield, D. L. (2016). Effect of dietary
components on the gut microbiota of aquatic animals. A never-ending story?
Aquaculture Nutrition, 22: 219-282.

59

Roeselers, G., Mittge, E. K., Stephens, W. Z., Parichy, D. M., Cavanaugh, C. M.,
Guillemin, K., & Rawls, J. F. (2011). Evidence for a core gut microbiota in the
zebrafish. The ISME Journal, 5(10): 1595–608.
Romero, J., & Navarrete, P. (2006). 16S rDNA-based analysis of dominant bacterial
populations associated with early life stages of coho salmon (Oncorhynchus
kisutch). Microbial Ecology, 51(4): 422–430.
Romero, J., Ringø, E., & Merrifield, D. L. (2014). The Gut Microbiota of Fish. In
Merrifield, D. & Ringo, E., editors Aquaculture Nutrition: Gut Health, Probiotics
and Prebiotics, 75-100, John Wiley & Sons, Ltd., Chichester, UK.
Saffron, I. (2004). The decline of the North American species. In Lebreton, G. T. O.,
Beamish, F. W. H. & Mckinley, R. S., editors Sturgeon & Paddlefish of North
America, pp: 1–21). Netherlands: Kluwer Academic Publishers.
Shade, A., & Handelsman, J. (2012). Beyond the Venn diagram: The hunt for a core
microbiome. Environmental Microbiology, 14(1): 4–12.
Sloan WT, Lunn M, Woodcock S, Head IM, Nee S, Curtis TP. (2006). Quantifying the
roles of immigration and chance in shaping prokaryote community structure.
Environmental Microbiology, 8:732–740.
Smit, J. (1911). The Biology of Fermenting Sarcina. Journal of Pathology, 35: 455-468.
Stackebrandt, E., Lang, E., Cousin, S., Pauker, O., Brambilla, E., Kroppenstedt, R. and
Lunsdorf, H. (2007) Deefgea rivuli gen. nov., sp nov., a member of the class
Betaproteobacteria. International Journal of Systematic and Evolutionary
Microbiology, 57: 639–645.
Stephens, Z. W., Burns, A. R., Stagaman, K., Wong, S., Rawls, J. F., Guillemin, K., &
Bohannan, B. J. M. (2015). The composition of the zebrafish intestinal microbial
community varies across development. The ISME Journal, 1–11.
Tetlock, A., Yost, C. K., Stavrinides, J., & Manzon, R. G. (2012). Changes in the gut
microbiome of the sea lamprey during metamorphosis. Applied and Environmental
Microbiology, 78(21): 7638–44
Vadstein, O., Bergh, Ø., Gatesoupe, F.-J., Galindo-Villegas, J., Mulero, V., Picchietti, S.,
Scapigliati, G., Makridis, P., Olsen, Y., Dierckens, K., Defoirdt, T., Boon, N., De
Schryver, P., Bossier, P. (2012). Microbiology and immunology of fish larvae.
Reviews in Aquaculture, 5, S1–S25.
Vedrasco, A., Lobchenko, V., Pirtu, I., & Billard, R. (2002). The culture of live food for
sturgeon juveniles, a mini review of the Russian literature. International Review of
Hydrobiology, 87(5–6): 569–575.

60

Vellend, M. (2010). Conceptual synthesis in community ecology. The Quarterly Review
of Biology, 85(2):183–206.
Venkataraman, A., Bassis, C. M., Beck, J. M., Young, V. B., Curtis, J. L., Huffnagle, G.
B., & Schmidt, T. M. (2015). Application of a neutral community model to assess
structuring of the human lung microbiome. mBio, 6:e02284-14.
Verschuere, L., Rombaut, G., Sorgeloos, P., & Verstraete, W. (2000). Probiotic Bacteria
as Biological Control Agents in Aquaculture, 64(4): 655–671.
Wang, Y. L., Binkowski, F. P., & Doroshov, S. I. (1985). Effect of temperature on early
development of white and lake sturgeon, Acipenser transmontanus and A.
fulvescens. Environmental Biology of Fishes, 14: 43-50.
Wickham, H. (2009). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag,
New York.
Wilson, J. A., & Mckinley, R. S. (2004). Distribution, Habitat, and Movements. In
Lebreton, G. T. O., Beamish, F. W. H. & Mckinley, R. S., editors Sturgeon &
Paddlefish of North America, pp: 40–72. Netherlands: Kluwer Academic Publishers.
Wong, S., Stephens, W.Z, Burns, A. R., Stagaman, K., David, L. A., Bohannan, B. J. M.,
Guillemin, K., Rawls, J. F. (2015). Ontogenetic differences in dietary fat influence
microbiota assembly in the zebrafish gut. mBio, 6(5), 1–9.
Wu, H. J., & Wu, E. (2012). The role of gut microbiota in immune homeostasis and
autoimmunity. Gut Microbes, 3(1): 4–14.
Yan, Q., van der Gast, C. J., & Yu, Y. (2012). Bacterial community assembly and
turnover within the intestines of developing zebrafish. PLoS ONE, 7(1).
Yatsunenko, T., Rey, F. E., Manary, M. J., Trehan, I., Dominguez-Bello, M. G.,
Contreras, M., Magris, M., Hidalgo, G., Baldassano, R. N., Anokhin, A. P., Heath,
A.C., Warner, B., Reeder, J., Kuczynski, J., Caporaso, J. G., Lozupone, C. A.,
Lauber, C., Clemente, J. C., Knights, D., Knights, R.,Gordon, J. I. (2012). Human
gut microbiome viewed across age and geography. Nature, 486(7402):222-227

61

CHAPTER 2: COMPOSITIONAL DYNAMICS OF LAKE STURGEON GUT
MICROBIOMES ASSOCIATED WITH DIETARY TRANSITION DURING
EARLY ONTOGENETIC STAGES
ABSTRACT
Compromised nutritional status associated with dietary transitions in aquaculture
operations, especially during early ontogenetic stages can have significant impacts on the
growth and survival rates in numerous fish species. In this study, we performed a 36-day
feeding experiment using 2 treatments involving control and food-transitioned groups to
quantify the impact of dietary switch on both physiological response and gut microbiome
composition in lake sturgeon (Acipenser fulvescens) larvae. We also investigated
differences in taxonomic composition of protease-producing bacteria using culture-based
methods and Sanger sequencing. Sturgeon that were transitioned from brine shrimp to
frozen bloodworms differed significantly in terms of growth (total length=p < 0.001;
weight = p < 0.0001) and survival rates (p < 0.01) based on mixed model ANOVA
statistical test. Individuals in the transition group probably experienced starvation relative
to fish in the control group that were fed only brine shrimp throughout the experiment.
Dietary treatment affected the taxonomic composition and diversity of gut microbial
communities as well, although there are successional changes observed in both groups
across development periods. Massively parallel sequencing using a region of the 16S
rRNA gene revealed that the majority of microbial taxa obtained from the spiral valve of
lake sturgeon larvae in the transitioned group were dominated by members of the phyla
Proteobacteria at 36 days post feeding (dpf), whereas gut microbial communities of fish
in the control group were dominated by taxa from the phyla Firmicutes. Alpha and beta
diversities were significantly different between dietary treatments following introduction

62

of bloodworms to members of the transitioned group. We identified 98 proteaseproducing taxa in fish using Sanger sequencing. Analyses revealed that isolates were
mostly from genera Pseudomonas and Aeromonas. However, fish that received
bloodworms included bacterial taxa from a greater phylogenetic diversity of genera
associated with family Enterobacteriaceae. Collectively, data suggest dietary transition
affects the composition of intestinal microbiomes of lake sturgeon larvae. Data will be of
assistance to hatchery managers when considering feeding practices that are believed to
increase growth with establishment of healthy, normal gut microbiota during larval
stages. This study demonstrates the utility of using microbiological interrogations of
microbial communities when characterizing dietary effects for important aquaculture
species.

63

INTRODUCTION
The growth of vertebrates occurs rapidly during early life stages. Growth and
development typically involves phenotypic changes that have a profound impact on
ecology, behavior, and physiology. In natural settings, fish experience size-related dietary
shifts as a consequence of size-specific changes in foraging abilities that enable
individuals to capture and consume progressively larger prey (Olson, 1996). Fish feed
either opportunistically or on a diurnal cycle (Wood & Bucking, 2010) in order to meet
the nutritional demands for growth.
Fish that were raised in artificial settings also experience shifts in diet as
aquaculture professionals attempt to accommodate hatchery management strategies with
rapidly changing physiological needs associated with (1) dietary requirements/nutritional
needs to maximize fish growth; (2) operational costs, (3) convenience associated with
regular food preparations (Hamlin et al. 2006). Ontogenetic food-switching and increases
in dietary requirements will directly influence amount of nutrient and food uptake in a
stable and constant supply of diet and dietary composition (Hamre et al., 2013).
High mortality during the period of first larval feeding has been reported as one of
the major bottlenecks in commercial fish production (Jobling, 2016; Li & Mathias, 1987;
M Hixson, 2014). The larval stage is characterized by rapid physiological and ecological
changes, and is associated with shifts from use of endogenous to exogenous nutrition
sources. At this time, larvae are most sensitive to environmental factors, particularly food
type and amount (Li & Mathias, 1987). Studies indicate that larvae receiving inadequate
food source or experiencing an imbalance in nutrients are often associated with poor
growth and malformation, which may result in decreased survival (Kjørsvik et al., 2011)

64

The rearing of larval fish of many important aquaculture species (e.g., salmonids,
sturgeon) depends on live food such as brine shrimp (Artemia salina), rotifers,
chironomids, and other zooplanktonic animals (Bjornsdottir et al., 2009; Léger,
Bengston, Sorgeloos, Simpson, & Beck, 1987; Ohs, Cassiano, & Rhodes, 2009;
Vedrasco, Lobchenko, Pirtu, & Billard, 2002). This dietary regimen must consider not
only the cost and labor associated with preparation of food, but also fish nutritional
requirements and suitability of diet that can be accepted, ingested, and digested by fish as
they grow (Jobling, 2016; Holt, 2011).
Approaches such as co-feeding or food transitioning are often employed to
improve larval growth and survival. Co-feeding (offering commercial or starter diets in
combination with live food), or transitioning (offering one food source then switching
gradually or instantaneously are two alternative feeding regimens that have been used
with numerous fish species) are widely used in commercial or conservation aquaculture
(Agh et al. 2012; Agh et al. 2013; Hamlin et al. 2006; Hamre et al., 2013). However,
previous studies have documented issues associated with providing suitable feed for
marine and freshwater larviculture (Agh et al., 2013; Bakke, Glover, & Krogdahl, 2010;
Naylor et al., 2009; Rabe & Brown, 2000; Zambonino Infante & Cahu, 2010). Issues are
often associated with low survival and growth.
A great deal of interest has focused on development of feeding strategies in fish
larviculture based on commercial diet as an economic alternative to live foods (Kolkovski
et al., 1997, Agh et al. 2013). However, inert larval feeds may not be a suitable substitute
for live food in terms of nutritional composition, palatability, digestibility or physical
characteristics of the dry feed, or inability to elicit larval feeding responses (Agh et al.,

65

2013; Holt, 2011; Jobling, 2016; Langdon & Barrows, 2011; Ohs et al., 2009; Oie et al.,
2011; Yufera, 2011; Zambonino Infante & Cahu, 2010). Therefore, approaches using live
prey such as brine shrimp, copepods, rotifers, and other aquatic macroinvertebrate like
bloodworms (Family: Chironomidae) are used to raise larval aquatic species.
The nutrient environment greatly influences the microbial communities
(microbiota) inside the gastrointestinal tract (GI) of vertebrates (Ursell et al., 2012). Diet
composition is also associated with gut microbiota composition in fish (Ringø et al.,
2015). Fish rely heavily on gut microbiota for many physiological functions including but
not limited to growth, digestion, nutrient production and absorption, protection of animal
against pathogens and immune system function (Llewellyn et al., 2014; Wu & Wu, 2012;
Hooper et al., 2012; Cain & Swan, 2010; Gómez & Balcázar, 2008; Macpherson &
Harris, 2004). Thus, the presence of a healthy gut microbiota is essential to normal host
functioning.
Any disruption in food and nutrient supplies may alter microbiota composition.
For instance, fish hosts that experience food deprivation in the wild or cultivated fish that
failed to convert during a food transition undergo a phase that represents an ‘energy
crisis’ to themselves as well as to gut microbiota due to absence or reduction in the
availability of nutrients within the fish host gut ecosystem (Kohl et al., 2014; McCue,
2012). Previous studies indicate alterations of microbial community composition are
linked to the quantity and composition of diets. However, questions such as what
microbial populations were diminished during the phase or remain unanswered due to
lack of studies on microbial community dynamics over the course of prolonged food
deprivation.

66

Sturgeons are important aquaculture species worldwide that are raised for
commercial and conservation purposes. Lake Sturgeon Acipenser fulvescens is an
imperiled species in the Laurentian Great Lakes Region (Peterson et al. 2007), and
populations are supplemented through conservation aquaculture programs by stocking
hatchery-reared juveniles (Crossman et al. 2011). Hatchery rearing of this species
presents difficulties, primarily associated with high mortality during early ontogenetic
stages due in part to poor feeding performance (Czeskleba et al., 1985; Hung & Deng,
2002; Klassen & Peake, 2008).
Low larval survival is linked to nutritional regimes associated with diet
formulation, feeding schedule, food presentation and preference (Mim, Lazur, Shelton,
Gomelsky, & Chapman, 2002; Vedrasco et al., 2002). In the hatchery production of
sturgeon, brine shrimp Artemia nauplii and bloodworms are commonly chosen as live
foods for feeding fish larvae and fry due to their high nutritional values (protein content),
digestibility and high conversion efficiency (Agh et al., 2012; Agh et al., 2013; Ceskleba
& Avelallemant, 1985; DiLauro et al., 1998; Vedrasco et al., 2002; Volkman et al.,
2004). Artemia is easily cultured and harvested on demand from dry and storable dormant
cysts (Das et al., 2012; Léger et al., 1987; Oie et al., 2011). Meanwhile, bloodworms
larvae have been shown to be a major component of lake sturgeon diet in wild and are
deemed suitable for hatchery production of bottom feeders (Das et al., 2012; Volkman et
al., 2004).
Newly hatched Artemia are typically fed to sturgeon at the onset of exogenous
feeding, yet the preparation of this food is costly and labor intensive. As fish grow, larvae
transition to frozen bloodworms (Diptera: Chironomidae). This dietary transition usually

67

takes place 2– 4 weeks post exogenous feeding (Klassen & Peake, 2007). Transitioning
larvae from brine shrimp to bloodworms reduces the harvesting effort associated with
Artemia preparation, and decreased variability in diet nutritional quality (Klassen &
Peake, 2008; Volkman et al., 2004; Léger et al., 1987). In addition, comparative studies
conducted by Volkman et al. (2004) showed that fish fed a greater proportion of
bloodworms compared to brine shrimp alone experienced higher growth rates. However,
experimental data are lacking for lake sturgeon growth performance during periods of
dietary transition, specifically effects of the diet transition on gut microbial community
composition.
The main objective of this study was to document the composition of gut
microbial communities in lake sturgeon during early ontogenetic stages associated with
feeding transition and to quantify fish growth. We further compared gut microbial
communities as fish were raised using different diet regimes and additionally assessed
culturable proteolytic bacteria by culture-based methods to identify associations of diet
treatment on the diversity of proteolytic species involved in digestion. We hypothesize
that gut microbial community composition of treated fish at times before, during, and
after diet transitions will differ relative to control fish (fish fed only Artemia).
Documentation of gut microbial communities will contribute insight to key features of
host-microbes relationships in the larval lake sturgeon gut. A successful dietary transition
would lead to increased growth and survivability of cultivated fish.

68

METHODOLOGY
Fish husbandry and feeding experiment
The feeding experiment was conducted at the Black River Sturgeon Rearing Facility
managed by the Michigan Department of Natural Resources (MDNR) and Michigan
State University (MSU) in Onaway, MI, USA using approved Institutional Animal Care
and Use Committee (IACUC) protocols. Lake sturgeon larvae were produced from a
single female and one male adult lake sturgeon from gametes collected during the 2015
sturgeon-spawning season (May 3rd) from the Upper Black River, in Cheboygan County,
MI, USA. Gamete collection and eggs fertilization method were described by Crossman
et al. (2011) and Bauman et al. (2015), respectively. Full-sib individuals were used to
reduce potential variability in microbiota associated with different host genotypes. All
individuals were raised using the same conditions until the feeding experiment began.
Fertilized eggs were maintained in Aquatic Eco-systems (Pentair) J32 Mini-Egg hatching
jars during incubation to simulate stream water flow. Egg mortalities were monitored
daily and removed. Upon hatching (May 9, 2015), yolk sac fry were moved to 3.0 L
polycarbonate aquaria (Aquatic Habitats). Fry were later assigned randomly to six 3.0 L
aquaria of density fifty fish per tank at eight days post hatch (dph; Bauman et al. 2015).
Water supplied to hatching jars and aquaria were taken from the Upper Black
River and filtered through 100-micron and 50-micron filters to remove large sediments
throughout the experiment. Water temperature was monitored hourly using a YSI
ProODO Optical DO-Temp meter. Average daily water temperature ranged from 11.8 ˚C
to 19.7 ˚C, with the mean (±SD) being 16.4 ± 1.8 ˚C over the duration of this study. The
mean (±SD) flow rate for each 3.0-L aquaria was approximately 440 mL/min (8.8 aquaria

69

cycles/hour). A 9 h light, 15 h dark cycle was maintained using fluorescent lights. At nine
dph, aquaria were randomly assigned to two treatment groups: Control (CR) and Treated
(TR) with three replicates each (n = 3 experimental units). Larvae in the CR treatment
were fed live Artemia nauplii throughout the duration of the study (five weeks).
Meanwhile, larvae in the TR treatment were fed Artemia nauplii during the first two
weeks post exogenous feeding (pef; or until 14 days post-feeding, dpf). Individuals were
gradually transitioned to frozen, commercially grown bloodworms during week three pef.
During week four until the end of the experiment, all individuals in the TR treatment
were fed only bloodworms. The feeding regime was formulated based on Deng et al.,
(2003) and Bauman et al. (2016) where the amount of food offered varied by week. In
weeks 1 and 2, 26% dry body weight (dbw) was offered; weeks 3 and 4, 13% dbw; week
5, 11% dbw. Each day, fish were fed these amounts based on dbw. On the initial
transition day, TR fish were fed 90% dbw brine shrimp and 10% dbw bloodworms. On
subsequent days, ratios for Artemia and bloodworms were as follows: day 2, 80% and
20%; day 3, 60% and 40%; day 4, 50% and 50%; day 5, 40% and 60%, day 6, 30% and
70%, and day 7, 15% and 85%. Fish were fed an equal amount of food three times daily
(0900, 1300, and 1700) while water remained flowing, simulating stream flow.
Brine shrimp cysts, Great Salt Lake strain, were purchased from Brine Shrimp
Direct© and were cultured following manufacturer's protocols. Frozen one-inch
bloodworms, harvested from ponds in northern China, were purchased from Brine
Shrimp Direct©. Bloodworms were rinsed with ground water and mildly chopped using
the pulse function on an electric blender before feeding.

70

Beginning nine dph, mortality per aquarium was recorded daily. Fish were batch
weighed by replicate once a week to measure growth and alter the amount of food
allocated to each aquarium in the following week. Each replicate was weighed separately.
All fish were collected in a dip net and water was removed by dabbing the net on the
dried paper towel several times before weighing to remove excess water. Wet weight of
Artemia was calculated from dry weight at dry weight = 0.1767 [sieved wet weight] 0.0541 (Bauman et al. 2016). Wet weight of bloodworms was calculated from dry weight
at dry weight = 0.0832 ·[wet weight] + 0.0239 (Scribner, unpublished data).
Sample collection
Sampling occurred at three ontogenetic stages of development: - two weeks (14 dpf),
three weeks (21 dpf), and five weeks (36 dpf) post exogenous feeding. Prior to sampling,
the food was not administered to the fish for 18 hr. Five fish were collected per replicate
and treatment group and anesthetized with MS-222 (Sigma-Aldrich, St Louis, MO, USA)
in a petri dish with a ruler. Fish were digitally photographed to obtain total length
measurements before being preserved in liquid nitrogen for enzymatic studies. Digital
images were later analyzed using Image J software (NIH Image). Another five fish were
sampled for microbial community interrogation at each of the three stages. Fish were
euthanized with an overdose of MS-222. Individuals were subsequently placed in
microcentrifuge tubes and stored at room temperature in 80% ethanol. During the fish
sampling, water samples were also collected. 500mL of stream water was filtered using a
0.22 um 47 mm filter membrane (Sterlitech®) using a hand-pump. The membrane was
placed in a 50 mL tube and preserved with 80% ethanol. All fish and water samples were
stored until dissection and bacterial DNA extraction was performed.

71

Fish dissection
The distal gut (spiral valve) of each sturgeon larval gastrointestinal tract was recovered
from fish following aseptic techniques. The distal gut was defined as the section that
begins at the intestine until reaching the spiral valve. The spiral valve serves as the
primary region of digestion and absorptive function (Buddington & Christofferson, 1985;
Callman & Macy, 1984), and thus it is believed to provide an area of abundant nutrients
where microbial communities can flourish. Exterior surfaces were swabbed with 100%
ethanol before dissections of the whole digestive tract using sterile instruments.
Dissections were performed with slight modification as previously described by
(Milligan-Myhre et al., 2011). The intact alimentary tracts were removed from the fish
body cavities, and the excised gut was immediately transferred into filtered-sterilized
80% ethanol solution for DNA isolation. Due to the small size of each larval gut, a
composite of four guts from four individuals was grouped for each replicate (or tank)
within CR and TR groups at each of the three-time points.
Isolation of bacterial culture & extracellular protease screening
At the end of the feeding experiment, four fish were sampled from replicate randomly
chosen from each of the treatment groups (CR and TR). Fish were euthanized and
immediately dissected following the aseptic dissection procedure previously described.
The freshly dissected gut was transferred to a tube containing 10ml of sterile LuriaBertani (LB) broth and transported to Michigan State University for further processing.
To prepare the intestinal homogenate, guts from two larvae were transferred into a
separate microcentrifuge tube, containing 250 ul of sterile LB. The guts were
subsequently homogenized using a sterile glass rod. Homogenates of guts from another

72

two larvae, originating from each replicated CR and TR groups, were pooled to create a
composite mixture of gut microbial communities for each group.
Homogenate samples were serially diluted in sterile LB up to 1:108. Once the
dilution was made, an aliquot of 0.1ml of each dilution was plated on each of three
replicate plates of nutrient LB agar using the spread plate technique. Plates were
incubated at 25oC for 48-72 hr. Following incubation, all colonies were counted, and their
characteristic morphologies were documented. Each isolated colony was subcultured in a
microtiter plate well containing 150ul sterile LB media, incubated overnight, and stored
at -80oC in 15% glycerol concentration as a stock bacterial culture for further use.
Isolates of pure colonies were subjected to a plate assay to screen for extracellular
microbial protease activities (Tetlock et al., 2012; Kazanas, 1978). Milk protein agar
plates were prepared in duplicates as follows: 20% (w/v) skim milk solution and 2x
purified LB agar was prepared separately in deionized water and autoclaved at 121oC for
15 minutes. The milk solution was then added to agar to give a final volume of 10%
(w/v) of milk protein agar. The mixture was kept at 55oC before poured into plates. All
isolates were cultured on the agar surface using 96 solid pin multi-blot replicators and
incubated at room temperature for the first 24hr and transferred into 30oC for the next 24
hr.
The zone of clearing around an inoculated colony indicated positive protease
activity, these colonies were scored as either 0 (negative) or 1 (positive). The clearing
phenotype was recorded and photographed. Protease-producing isolates were subjected to
Sanger sequencing for the v4 region of the16S rRNA gene at the Michigan State
University Research Technology Support Facility, RTSF (East Lansing, MI, USA) for

73

further identification. Sequencing of 600bp 16SrRNA via Sanger was preceded to further
identify these potential protease producers.
Sanger sequencing data were analyzed using Ribosomal Database Project (RDP,
www.rdp.cme.msu.edu) pipeline. Sequences were first converted into FASTA format and
manually trim to remove the ambiguous bases. These sequences were first aligned and
classified with RDP’s aligner and RDP Classifier (Wang et. al, 2007). Sequences that
included too many ambiguous positions were not included for the alignment.
Phylogenetic trees were then constructed using RDP’s Tree Builder with bootstrap
confidence estimates using weighted version of the Neighbor-Joining method (Saitou &
Nei, 1987) based on a distance matrix calculated using the Jukes-Cantor model (Jukes &
Cantor, 1969). The branch supports were estimated in a bootstrap analysis based on 100
pseudoreplicates for NJ trees bootstrap estimates were generated and displayed at the
branches of the majority consensus tree.
DNA extraction and 16S rRNA amplicon sequencing
Gut microbiota from lake sturgeon larvae was surveyed using high-throughput
sequencing of the v4 region of the 16S rRNA gene. Each gut sample was first centrifuged
for 15 min at 4oC to pellet bacteria before DNA was extracted. The MoBio PowerSoil®
DNA Isolation Kit (Carlsbad, CA, USA) including a bead-beating step was used
following protocols for low-biomass samples, as suggested by the manufacturer. The
integrity of each DNA sample was assessed based on amplification of 1.4k bp of the 16S
rRNA gene (amplicon based on 27F and 1389R primers) followed by gel agarose
electrophoresis. DNA concentrations were quantified using a Microplate
spectrophotometer (BioTek®, Winooski, VT, USA).

74

Twenty-seven DNA samples (including two control samples) were sequenced at
Michigan State University Research Technology Support Facility, RTSF (East Lansing,
MI, USA). All sequencing procedures, including the construction of the Illumina
sequencing library, emulsion PCR, and MiSeq paired-end sequencing v2 platforms of the
V4 region (~250bp; primer 515F and 806R) followed standard Illumina (San Diego, CA,
USA) protocols. The Michigan State’s Genomics RTSF
(https://rtsf.natsci.msu.edu/genomics/) provided standard Illumina quality control,
including base calling by Illumina Real Time Analysis v1.18.61, demultiplexing, adaptor
and barcode removal, and RTA conversion to FastQ format by Illumina Bcl2Fastq v1.8.4.
Sequence processing
Sequence data were processed using default sequencing data analyses pipeline and
computing workflow. Briefly, paired-end sequence merging, quality filtering,
“denoising”, singleton-sequence removal, chimera checking, taxonomic assignments and
Operational Taxonomic Unit (OTU) selection was conducted based on methods
implemented by program mothur v.1.36.1 (Kozich, Westcott, Baxter, Highlander, &
Schloss, 2013). A reference-based OTU clustering and the taxonomic assignment was
performed using SILVA-based microbial reference database file provided in mothur to
cluster sequences defined with 97% identity. To minimize effects of under sampling
while maintaining as broad dataset as possible, final OTUs were rarefied to a depth of
9151 sequences per sample. Two DNA samples with low sequence depth were discarded
in downstream analyses. Rarefaction analyses were performed to evaluate the sampling
coverage for each sample based on selected sequence depth.

75

Morphometric data and survival analyses
The effects of food transition on larval growth were quantified based on total body length
(TL) and batch (replicate/tank) weight (WT). Each of the 3.0-L aquaria represented an
experimental unit, and mean weight (g ± SD) and length (mm ± SD) for three replicates
for each treatment group (CR vs TR) at each sampling time are displayed in Table 2.1
and Figure 2.1. Mean TL and mean WT per fish that were repeatedly measured over time
were analyzed using a mixed model approach. In this approach, we compared the mean
differences of TL and WT between two feeding treatments (CR vs TR; i.e “feeding” is a
“between-subject” factor) over sampling points/stages (i.e., “time” was the “withinsubject” factor).
We wished to quantify mean TL and WT per fish as a function of time and as a
function of feeding treatment (CR vs TR), as well as to determine potential interactions
among stages and treatments. While larval total body length was measured based on three
tank replicates at three sampling points, larval body weight was based on three replicates
of each feeding treatments measured weekly throughout the experimental duration.
Survival was estimated as the mean daily proportion of larvae surviving from week 0
through week 5 pef in R using survival package (Therneau & Grambsch, 2000) and
survival curve was generated survfit function. Proportional survival analyzed using two
samples log-rank test using the survdiff function. This function was based on a ChiSquare distribution that evaluates whether fish from different feeding treatments are
coming from the same distribution or two different distributions. All statistical analyses
were conducted using programming and statistical software, R (version 0.98.978). A pvalue below 0.05 was considered statistically significant.

76

Analyses of bacterial community profiles and ecological statistical analyses
2.1 Alpha diversity
All measures of microbial community diversity including Inverse Simpson (1/D)
diversity indices and OTUs richness of each sample were calculated from the sequence
data within program mothur. To test for significant differences in diversity indices
between CR and TR treatment groups, we first employed a log transformation of Inverse
Simpson indices and observed a number of taxa followed by two-way ANOVA using the
statistical software, R (version 0.98.978) base package. The test was followed by posthoc
Welch two samples t-tests. p-values below 0.05 indicate significant differences in
pairwise comparisons.
Relative abundance and identity of all phyla in all fish gut and water-associated
microbial communities across sampling times were determined using packages dplyr and
reshape2 in program R.
2.2 Beta diversity
We used several packages implemented in R to estimate comparative (beta [β]) diversity
measures and ecological statistics at the OTU level. Briefly, we used the vegan package
(Oksanen, 2015) to generate Bray-Curtis (BC) distance characterizing differences in
microbial community composition. We used the cmdscale function to perform Principle
Coordinate analyses (PCoA) ordination of community composition differences based on
BC distance (Bray & Curtis, 1957). The ggplot and ggplots2 packages (Wickham, 2009)
were then used to create ordination plots to visually compare gut community composition
with aquatic community composition as a function of different treatments and among
sampling periods based on the principal coordinate with the three largest eigenvalues.

77

Multivariate hypotheses testing to quantify differences in community composition
among samples collected at different times and from different feeding treatments were
performed using the adonis function (Oksanen, 2015) in program R. Differences between
locations of the centroids of sampling groups for treatments and time periods were based
on Permutational Multivariate Analyses of Variance (PERMANOVA) using BC
resemblance matrices (2013, 2006; 2001a). Under the null hypotheses, food treatments
were not expected to significantly affect fish gut community taxonomic composition.
This test was employed because of the non-parametric and skewed nature of microbial
ecology data distribution. Then, object scores along the principal coordinate with the
three largest eigenvalues (i.e., eigenvectors associated with corresponding eigenvalues)
were correlated with object scores along each original variable’s axis (i.e., relative
proportions of measured taxa/OTU) to measure the OTUs’ contribution to a given PCoA
axis (Legend & Legendre, 2012) using the corr function in program R.
Inferring gut-associated microbial communities function
Metagenomic analyses of 16S rRNA sequencing data and reference genomics databases
were used to predict the functional roles of gut microbiota associated with treated fish
(TR) and control fish (CR) groups using program PICRUSt (version 1.1.0; Langille et al.,
2013) We imputed the putative function of lake sturgeon gut using catalogues of
annotated genes within a known sequence database (Kyoto Encyclopaedia of Genes and
Genomes; KEGG) based on the abundance of taxa with known function using taxonomic
identification established from our 16S rRNA survey. With PICRUSt, we calculated a
Nearest Sequenced Taxon Index (NSTI), which measures how closely related the average
16S rRNA sequence in the sample was to a sequenced genome. When this index is low (<

78

0.05), PICRUSt is likely to perform well in predicting the genomes of the organisms in
an environmental sample. High scores (> 0.15) mean few related references are available
and prediction will be of low quality (Langille et al. 2013).
Briefly, reference-based OTU clustering and taxonomic assignments were
performed using selected marker gene identifiers in the Greengenes database and
converted into a biom file. The file was transferred into the PICRUSt program, and
functional predictions were carried out based on information of the relative abundances
of those OTUs across samples using an evolutionary model. The OTU table for each
sample with associated Greengenes identifiers was later normalized based on organisms
predicted 16S gene copy number using normalize_by_copy_number.py script. Functional
roles were predicted by searching for pre-calculated genome content for each OTUs using
predict_metagenomes.py script. Annotations of predicted function were applied and
summarized using the KEGG Orthology (KO) classification schemes using the
categorize_by_function.py script, all included in PICRUSt. The program generated
profile tables consisting of annotated gene functions along with abundance for each
sample in the OTU table (Langille et al. 2013) to provide a baseline to infer the
functional attributes of observed taxa. In addition, PICRUSt calculated the NSTI to
quantify dissimilarity between reference genomes and the predicted metagenome
presented here. The graphical representation and Welch’s t-test were performed both in
program STAMP (Parks, Tyson, Hugenholtz, & Beiko, 2014) and R to quantify
significant differences in the mean proportion of sequence associated with predicted
functions.

79

RESULTS
Growth performance and survival
The proportion of fish surviving in each diet treatment was statistically different between
CR and TR treatments (Fig. 2.1). Larval fish survival among lake sturgeon in the TR
treatment from Artemia to bloodworms was significantly lower with nearly 50%
mortality relative to fish in the CR treatment that mostly survived by the end of 5-week
study (Chi square value, χ2 = 85.1, p-value < 0.01, df =1). Mortality peaked from 30 to 35
dpf.
Analyses of morphometric data (TL and WT; Fig. 2.2a and 2.2b) were achieved
by comparing the mean weight of fish (g ± SD) and mean length (mm ± SD) between
treatments across time periods. There were significant interactions between sampling
time and dietary treatment, influencing the mean length among replicates (df=2, F=
99.62, df = < 0.0001, Table 2.1a). At 14 dpf and 21 dpf, no significant differences were
detected between mean lengths from fish in both treatments (14 dpf: CR 30.84 ± 1.00
mm, TR 31.13 ± 1.05 mm, p-value = 0.66; 21 dpf: CR 35.59 ± 1.65 mm, TR 34.63 ± 1.36
mm, p-value = 0.32). However, mean length for TR (33.15 ± 2.36 mm) was recorded to
be significantly lower (p-value < 0.001) than CR (52.08 ± 3.64 mm) as fish reach 36 dpf.
Similarly, a strong interaction existed between time (sampling week) and
treatment, influencing weight (df=3, F= 111.5, p= < 0.0001, Table 2.1b). Comparison
between dietary treatments for each week indicated mean weight for weeks 5 and 6 were
significantly different (week 5: CR 0.267 ± 0.04 g, TR 0.101 ± 0.01 g, p-value < 0.05;
week 6: CR 0.463 ± 0.05 g, TR 0.104 ± 0.01 g, p-value < 0.01) whereas for week 3 and
week 4, they were not different (CR 0.111 ± 0.01 g, TR 0.105 ± 0.01 g, p-value =0.459;

80

week 6: CR 0.157 ± 0.02 g, TR 0.121 ± 0.00 g, p-value=0.058). Measurements at the first
and second week were not included due to the lack of data from all aquaria replicates.
Characterization of diversity and proportion of microbial phyla in gut community
composition of lake sturgeon
At 14 dpf, gut microbial community samples of twenty-four fish from all six replicates in
CR and TR treatment groups were collected and sequenced. Comparisons at the level of
phyla indicated that six phyla were present in both feeding groups. Although the relative
abundance of these phyla differed between feeding groups, all together those phyla were
collectively dominate, including more than 90% of total gut microbial taxonomic
composition; Proteobacteria (mean TR 53.7%, mean CR 65.6%), Firmicutes (mean TR
14.5%, mean CR 27.1%), Actinobacteria (mean TR 8.8%, mean CR 5.1%),
Acidobacteria (mean TR 0.05%, mean CR 0.39%), Bacteroidetes (mean TR 12.9%, mean
CR 0.4%), and Verrucomicrobia (mean TR 6.5%, mean CR 0.2%). The differences in
mean proportions of these phyla are shown in Figure 2.3a. Each feeding group also had
three unique phyla, yet these taxa were present in low proportions (~ 1% or less).
Once bloodworms were added to the TR feeding treatment’s regime during the
transition week (21 dpf), we observed changes in microbial community composition in
samples in the TR group. Greater numbers of phyla were detected in TR fish samples
(Chlorobi, Fusobacteria, Lentisphaerae, Nitrospira, OD1, OP11, SR1, Synergistetes,
TM7) and these phyla were absent in fish the CR treatment that were fed only Artemia
(data not shown). We also observed an increased proportions of Acidobacteria (mean
relative abundance increased from less than 1% to 5.9%) and unclassified phyla (mean
relative abundance increased from 2.5% to 8.1%) as fish grew from 14 dpf to 21 dpf in

81

the TR treatment. Interestingly, in the same feeding group, we documented declines in
abundance of Firmicutes (14 dpf: 15%, 21 dpf: 6%), while Firmicutes proportions
remained relatively stable in CR treatment (14 dpf: 27.2%, 21dpf: 20.2%). At 21 dpf, the
gut community composition for fish in the CR group was dominated by three major phyla
(Firmicutes, Actinobacteria, Proteobacteria) totaling more than 95% of the total number
of sequences. At the end of the experiment (36 dpf), the gut microbial communities from
fish in both CR and TR groups were distinct. The Firmicutes phylum dominated
community composition from the CR group (93.5%), while gut communities of TR fish
were dominated by Proteobacteria (94.3%). Overall patterns of phyla present in the fish
gut contrast greatly with phyla detected in water samples during all three sampling
periods (Fig. 2.3b).
Measures of gut microbial community diversity including the Inverse Simpson
indice and taxa richness (number of taxa detected/observed taxa) indicated fluctuation of
index estimates across ontogenetic stages/sampling points (Figs. 2.4a, 2.4b).There was a
significant interaction of these stages (dpf) and dietary treatment influencing both indices
(logInverse Simpson: df = 2, F=4.483, p-value < 0.05; logTaxa richness: df = 2, F=18.53,
p-value <<0.001). Findings suggested that alpha diversity indices (Inverse Simpson and
taxa richness) of fish from diet treatments differ significantly during at least at one
sampling period.
Welch t-tests for two samples with unequal variances were later performed using
log-transformed diversity measures, to test for differences in mean diversity indices for
fish samples from different treatment at each sampling time. Statistical tests on both
measures of Inverse Simpson’s index and taxa richness at 14 dpf failed to reject the null

82

hypotheses at 0.05 significance levels (logInverse Simpson CR = 1.857, logInverse
Simpson TR = 2.448, p-value = 0.292; logTaxa rich. CR = 4.748, logTaxa rich. TR = 4.357,
p-value=0.236). However, log mean comparison between treatments at 21 dpf (logTaxa
rich. CR = 5.344, logTaxa rich. TR = 7.247, p-value < 0.05; logInverse Simpson CR =
2.609, logInverse Simpson TR = 3.777, p-value < 0.01) and 36 dpf (logTaxa rich. CR =
4.154, logTaxa rich. TR = 5.476, p-value < 0.001; logInverse Simpson CR = 0.419,
logInverse

Simpson TR = 0.333, p-value < 0.05) were significantly different (TR fish were

higher).
Association between gut microbial community composition and feeding treatments,
across sampling periods
Principal Coordinate Analyses (PCoA) ordination of differences in taxonomic
composition and relative abundance of the gut community was performed to visualize
relationships between those community composition associated with different feeding
treatments and times of collection. Microbial community composition of fish from both
CR and TR treatments were similar when fish were 14 dpf (Figs. 2.5a, 2.5b, 2.5c).
Subsequently, communities from fish in the transition group (TR) diverged while gut
communities of fish in control group remain relatively similar (Figs. 2.5a, 2.5b). Toward
the end of the experiment (36 dpf), gut community compositional differences between
feeding treatments resulted in treatments clustering further apart, although considerable
community composition overlap in fish and water communities was observed (see Fig.
2.5c). Statistical analyses of beta diversity indicated significant community compositional
differences among stages and between treatments (PERMANOVA test, pseudo-F =
2.928, R2 = 0.151, p < 0.001, df=2; Table 2.2).

83

We further examined variation in community composition by decomposing
treatment effects of diet from pervasive influences of developmental stages using linear
regression and least square mean analyses. At 14 dpf, no significant effect of treatment
was detected (F-statistics = 1.184, p-value: 0.3377). As diets of fish in the treated group
changed from brine shrimp to bloodworms at 21 dpf (F-statistics =140.2, p-value: < 0.01)
and continued through the end of the experiment at 36 dpf (F-statistics =1.213e+05, pvalue: < 0.001), communities differed significantly in taxonomic composition (Table
2.2b). PCoA axes-1 of all three sampling periods explained the most variation in
microbial communities representing the linear combination of all taxonomic composition.
Correlation analyses of the relative abundance of observed taxa/OTUs with
principal coordinate associated with three largest eigenvalues found the 17 taxa that
highly correlated (Pearson-correlation > 0.70) with associated eigenvector (Table 2.3).
For example, Otu011 (Unclassified Betaproteobacteria) strongly correlated with first
principal coordinate (Pearson-correlation = 0.91), Otu001 (Genus
Clostridium_sensu_stricto) with second principal coordinate (Pearson-correlation =
0.94), and Otu2 Otu010 (Genus Aeromonas) with the third principal coordinate (Pearsoncorrelation = 0.97).
Predicted functional roles of lake sturgeon gut microbiota
Functional inventories of microbiome genes, such as those involved in the metabolism of
macronutrients (carbohydrates, amino acid, lipids) can be predicted from bacterial species
assemblages using PICRUSt (Phylogenetic Investigation of Communities by
Reconstruction of Unobserved States). This study used the KEGG database to match the
chosen reference OTUs, focusing at the level 2 ( a function defined based on molecular

84

interactions, protein post-translational modifications). Calculation of NSTI from
PICRUSt across all 17 fish gut samples show that this metric fall within the range of
0.048 till 0.258, with an average of 0.22.
The program further inferred 37 gene families from the gut samples of fish from
both feeding treatments across sampling points (Figs. 2.6a, 2.6b, 2.6c). Of these 37 gene
families, the majority of the genes were associated with membrane transport (mean value
across sampling periods 13.58 ± 0.84% in CR-associated communities and 12.62 ± 0.81
in TR-associated communities, respectively), carbohydrate metabolism (11.75 ± 0.03% in
CR, 11.56 ± 0.66% in TR), amino acid (12.13 ± 0.45% in CR, 12.01 ± 0.70% in TR),
replication and repair (8.78 ± 0.64% in CR, 7.01 ± 0.29% in TR), and energy metabolism
(6.75 ± 0.25% in CR, 7.01 ± 0.29% in TR). The mean relative abundances of these
predicted gene families relatively stable across time. Of all 37 genes predicted, none of
means relative abundance differs significantly between treatments (data not shown here).
Activity-based screening for protease positive isolates
The screen for microbial taxa possessing genes constitutively expressing proteolytic
activity was based on the screens of zones of clearing characterizing pure colonies
isolated from gut samples. We collected a total of 288 pure colonies from individuals
from TR (160) and CR (128) diets cultured on the LB agar. Out of these pure colonies,
we then identified 52 positive-protease isolates for CR and 88 for TR when these
microbes cultured on skimmed-milk agar.

85

Sanger sequencing data generated for all 140 positive-protease isolates detected in
the fish gut from both TR and CR group. However, only 54 bacterial sequences from TR
and 44 sequences from CR were retained during sequence alignment and taxonomic
identification. The remaining taxa were filtered out due to presence of ambiguous base
positions in the FastA files. The Neighbor-Joining (NJ) tree generated for isolates from
fish in CR group (Fig. 2.7a) indicated the presence of genera Aeromonas and
Pseudomonas that are capable of hydrolyzing protein (casein) that was present in the
media. Meanwhile, in TR group, besides genera Aeromonas and Pseudomonas, six
additional genera in the family Enterobacteriaceae were also present; genus Buttiauxella,
Citrobacter, Enterobacter, Kluyvera, Leclercia, Yokenella (Fig. 2.7b). Species displayed
in both trees were based on top matches sequences selected based on Sab index higher
than 0.80 and to represent an extensive phylogenetic breadth of all species found during
alignment. Bacterial Thermotoga maritima was used as an outgroup.

86

DISCUSSION
Findings of significant differences in survival and low growth rates of TR fish observed
in this experiment relative to CR group prompted the feeding trial to be terminated at 36
dpf. Klassen et al. (2007, 2008) demonstrated successful transition of lake sturgeon when
they were switched from brine shrimp to black fly larvae several weeks after the fish
were first offered with brine shrimp. Nonetheless, transitioned fish in our study failed to
transition to bloodworms. Visual inspections on fish stomachs suggested that fish were
not eating bloodworm and were starving beginning at 21dpf. Lack of ingestion of
bloodworms may explain significant differences in mean total length and average fish
weight compared to fish in CR group. We suspect that frozen bloodworms that were used
in this study were inadequate to stimulate sensory organs and feeding response in fish. As
these inert feeds dissolve in culture water, the taste and smell might be affected, making
it no longer desirable to be ingested (Langdon & Barrows, 2011). Studies have shown
that live Artemia excretes metabolites that serve as strong feeding stimuli, perhaps
contribute to greater acceptance to fish larvae (Hamre et al., 2013; Yufera, 2011).
Our study presents comprehensive analyses of lake sturgeon gut microbiomes
using 16S rRNA amplicon based sequencing, offering an in-depth characterization of
longitudinal changes in microbial taxonomic composition over fish developmental stages
in response to diet treatments. We found that these fish exhibited distinct gut microbial
community at the end of the experiment (36 dpf) further supporting our hypotheses that
gut microbiota was affected by treatment. Phylum Proteobacteria dominated gut
microbial communities of fish from the TR group whereas fish in the CR group were
dominated by taxa in the phyla Firmicutes. Statistical analyses on measures of gut

87

microbial community diversity including the Inverse Simpson index and taxa richness
indicated that gut microbial communities in TR fish were significantly more diverse
compared to CR fish.
PERMANOVA test followed by linear regression analyses at each sampling
periods detected a significant influence of feeding treatment on gut community
composition of fish from both treatments occurred at 21 dpf and 36 dpf, but not at 14 dpf.
At the beginning of the experiment, gut microbial communities from all samples (CR and
TR) were expected to be similar since all fish received the same diet. The compositional
divergence in community composition and diversity was observed as soon as
bloodworms were introduced to the TR group. This divergence could either be attributed
to different diet associated with different microbiota or due to prolonged starvation as a
result of poor feed acceptance. We believe the latter is more likely.
Previous studies have indicated an alteration in gut community composition of
larvae that experience prolonged fasting or starvation. During the period of fasting, the
supply of nutrients to gut symbionts was significantly reduced, causing an “energy crisis”
(McCue, 2010, 2012). Fish that were starved also showed some intracellular degradation
and reduced endothelium, that subsequently reduce the mass tissue of host digestive
system – which in turn affected the taxonomic composition of microbiota (Kjørsvik et al.,
2011; McCue, 2012). It has been shown in the large intestine of Burmese python (Python
bivittatus), fasting was associated with increased abundance of genera Bacteroides,
Rikenella, Synergistes, and Akkermansia, and reduced overall diversity (Costello et al.
2010). Although our “starved” fish showed higher taxonomic diversity seen in alpha
diversity measure (Inverse Simpson and taxa richness), our findings reported the presence

88

of Synergistes in fish fed bloodworms at 21 dpf, and those communities were later highly
dominated by Proteobacteria at 36 dpf.
Our findings corroborate the results in a study performed by Kohl et al. (2014)
that reported a dramatic increase in relative abundance of Proteobacteria within
prolonged-fasting tilapia (Oreochromis niloticus) compared to well-nourished tilapia.
Results in another study by Xia et al. (2014) on Asian seabass (Lates calcarifer) was in
congruent with the Burmese python studies as they documented significant changes in
Bacteroidetes between control fish that were well-fed vs experimental fish that were
starved. In addition, we found a strong, positive correlation of the relative abundance of
genus Aeromonas with gut community composition of TR fish in our experiment. We
believe that though fish were not feeding on bloodworms, the presence of this food in the
water might indirectly affect the microbial community of water in the tank (De Schryver
& Vadstein, 2014; Skjermo & Vadstein, 1999). Bloodworms feed on detritus as well as
microalgae and they could possibly be a reservoir of zoonotic agents of several pathogens
such as Salmonella, V. cholera and could favor the transmission of infectious disease
organisms ( Rouf & Rigney, 1993; Broza & Halpern, 2001; Sharifian Fard et al., 2014).
Artemia is commonly fed to fish larvae in aquaculture settings. Studies have
identified Artemia as continuous, non-selective, phagotrophic filter-feeder that can ingest
food at a maximum rate. Due to this feature, Artemia could accumulate high bacterial
loads during the cultivation process that could later be transferred to fish or shellfish
larvae as they are eaten by these animals (Oie et al., 2011). It is possible that microbial
communities that established inside the gut of fish in the CR treatment were influenced

89

by microbes presence in Artemia as well as the continuous dispersal of communities from
surrounding water forming consortium of microbes that distinct from TR group.
Results of the PICRUSt predictive modeling from our study provide insights into
effects that realized changes in microbial community composition in nature and in
hatcheries may have on microbial community function that could have compensatory or
confounding and antagonistic effects to host well being (Langille et al. 2013). Gut
microorganisms play a number of functions in host physiology. For example, in
mammals, the gut microbiota plays an important role aiding the supply of alternative
energy sources, such as ketone bodies, when hosts are faced with fasting and starvation
(Crawford et al., 2009, Kohl et al., 2014). However, this prediction of gene family
function associated with our dataset has to be applied with caution due to high NSTI
metric quantified by PICRUSt. The high NSTI value indicates high compositional
discrepancies between our 16S metagenome dataset and reference genomes (Langille et
al. 2013). The fact that we are studying novel communities associated with the digestive
system of non-model and ancestral fish species that should have distinct taxa compared to
human gut microbial communities might attribute to this circumstance.
PICRUSt assignment of predicted metagenome content to Level 2 KO’s in our
data revealed that functional responses were generally consistent between treatments and
across time (Figs. 2.6a, 2.6b, 2.6c). Our study shows that the most abundant functional
categories were associated with carbohydrate metabolism, energy metabolism, amino
acid metabolism, membrane transport and replication and repair. This is consistent with
the general metabolic functions (such as carbohydrate, protein and amino acid
metabolism) that are essential for microbial survival (Mao et al. 2015). Studies by Xia et

90

al. (2014) also reported enriched microbial taxa (and genes) in Asian seabass (Lates
calcarifer) that have undergone starvation associated with functional categories including
carbohydrate transport and metabolism, inorganic ion transport and metabolism, and
amino acid transport and metabolism.
It is unknown whether the same functional groups of microorganisms perform
similar adaptive functions within other host taxa. While responses at the level of
microbial taxa to fasting vary across hosts, there may be certain microbial functions that
increase or decrease in abundance in fasted animals as shown by Xia et al. (2014). They
documented significant depletion of three categories (transcription, cell division and
chromosome partitioning, and replication, recombination and repair) while significant
enrichment of other six categories. On the other hand, we found no significant functional
differences between TR and CR group after correction for 37 multiple comparisons. As
we mentioned earlier, the different in the host taxonomic (sturgeon vs human) might
contribute to the non-significance findings. Sullam et al. (2015) also reported the same
finding when they performed multiple comparisons between difference ecotypes of
Trinidadian guppies. Nonetheless, almost all functional categories differed significantly
among when they compared across enterotypes. In their study, Sullam et al. (2015)
loosely defined enterotypes as differing community type presence in guppy gut systems
that have inherent functional different based on functional predictions.
Protease-producers found in the fish guts of both treatments were dominated by
members in genera Pseudomonas or Aeromonas, however, more genera were detected
fish from TR group. While we cannot completely rule out the possibility that culturebased screening methods could bias for identification, the inflated numbers of these two

91

genera could be attributed to the abundance of these bacterial taxa in the fish-surrounding
environment. Pseudomonas species are important decomposers of organic matter in soil,
water and food products; but several species also known as a pathogen in plants, animals,
and human (Palleroni, 1993). Pseudomonas also found to be commonly found as part of
fish intestinal and fish-egg microbial communities (Romero & Navarrete, 2006).
(Hoshino et al., 1997) has described a Pseudomonas sp. that exhibit protease activity at
low temperature isolated from the fish intestine.
Aeromonas, on the other hand, can be commonly isolated from various aquatic
environments and clinical tissue samples from human or animals and were also shown to
be present in both foods used in our treatments (Chironomids and brine shrimp) (Laviad
& Harpens, 2016; Austin & Allen, 1982). In addition, numbers of Aeromonas sp were
regarded not only as an important pathogen in poikilothermic animals, but also have been
as the etiologic agent causing many intestinal illnesses in human (Janda & Abbott, 2010).
Species like A. hydrophila ATCC 7966, and A. salmonicida can cause great loss in
commercially raised salmonids. Production of extracellular protease is one of the
virulence factors possessed by these.
Certain microbial taxa, such as Enterobacteriaceae are capable of initiating a
stringent response to reduce the rates of protein and other macromolecular synthesis by
decreasing rRNA synthesis in response to depletion of nutrition. This mechanism helps
cells to shut-down energy-draining activities and enter survival mode under poor growth
conditions (Yuan, 2006). This could explain the presence of more taxa under family
Enterobacteriaceae associated with fish from TR group as these fish experienced dietary
compromises during the transition, resulting in starvation, which eventually affects the

92

nutrient content needed by the microbial community in the gut. As a result, numbers of
microbial taxa thrived under this condition while others became significantly reduced.
Studies have listed both Pseudomonas and Aeromonas, as well as species within
the Enterobacteriaceae family as potential probiotics, and recent findings showed the
promising application of these taxa in aquaculture (for review see Balcázar et al., 2006;
Hoseinifar et al. 2014; Merrifield et al., 2010). Hence, given the importance of gut
microbial communities and potential application of advanced microbial-based strategies
in the artificial rearing of sturgeon and fish in general, further in depth studies should be
conducted to unravel the genotypic and functional diversity of these communities. The
efficiency feeding of live prey should also be evaluated and carefully monitored
especially during the first six weeks of exogenous feeding to larvae to ensure the health
and maximum growth.

93

APPENDIX

94

APPENDIX
Table 2.1: Split-plot ANOVA (mixed design ANOVA) table for (a) total length; (b)
average fish individual weight indicate source of variability between-tank replicate
associated with feeding treatment and within-tank replicate associated with sampling
times. A significant interaction was also observed between treatment and stages.
Table 2.1 (a)
Source

df

SS

MS

F

p-value

Trt group

1

191.94

191.94

64.41

0.001**

Error

4

11.92

2.98

Time

2

416.90

208.44

119.63

<0.0001***

Time*Trt Group

2

347.10

173.57

99.62

<0.0001***

Residuals

8

13.90

1.74

df

SS

MS

F

p-value

Trt group

1

0.1208

0.1208

121.8

<0.001**

Error

4

0.0040

0.0010

Week

3

0.1057

0.0352

101.5

<0.0001***

Week*Trt Group

3

0.1161

0.0387

111.5

<0.0001***

Residuals

12

0.0042

0.0004

Between subject (tank) effect

Within subject (tank) effect

Table 2.1 (b)
Source
Between subject (tank) effect

Within subject (tank) effect

95

Table 2.2 (a) PERMANOVA analysis indicates that variability among fish gut microbiota
differed significantly among developmental stages and between treatments. A significant
interaction was also observed between treatment and stages (PERMANOVA test pseudoF = 2.928, R2 = 0.087, p < 0.001; permutation=1000). (b) Goodness of fit from linear
regression model (R2), p-value, and Least-square means analyses performed on
significantly important PCo axes calculated separately for each stages period. PCo Axis 1
for stage 21dpf and PCo Axis 1 for stage 36dpf showed significant different across diet
treatment influencing the microbial community composition, but none of PCo Axis for
stage 14 dpf was found to be significant.
Table 2.2 (a)
Df

Sum

Mean

Sq

Sq

F-value

R2

Pr (>F)

Stages

3

2.710

0.903

3.518

0.273

< 0.001

Treatment

1

1.106

1.106

4.308

0.111

< 0.001

Stages:Treatment

2

1.504

0.752

2.928

0.151

<0.001

Residuals

18

4.623

0.257

0.465

Table 2.2 (b)
Linear regression model :

Least square means

Pco Axes ~ Treatment
Important axes
PCo1 14dpf (Before

Estimates

p-value

Control

Transitioned

0.321

0.338

-0.160±0.209

0.160±0.209

-0.754

<0.01*

0.302±0.040

-0.453±0.049

-0.990

<< 0.001**

0.330±0.002

-0.660±0.002

transition)
PCo1 21dpf (Transition
week
PCo1 36dpf (After
transition)

96

Table 2.3 List of taxa with taxonomic identification that show high correlation with
eigenvector of the first three Principal Coordinate axes associated with largest
eigenvalues
PCo. Taxa

Taxonomic identification/

Axes

Genera

2

Otu1

Correlation

0.94

Clostidium_sensu
stricto

3

Otu2

0.97

2

Otu10 Unclassified Clostridiceae

0.87

1

Otu11 Unclassified Betaproteobacteria

0.92

1

Otu12 Unclassfied Microbacteriaceae

0.84

1

Otu15 Unclassified Comamonadeceae

0.88

1

Otu16 Unclassified Actinomycetales

0.86

1

Otu17 Unclassified Comamonadeceae

0.83

1

Otu18 Unclassified Comamonadeceae

0.88

1

Otu19 Polynucleobacter

0.89

1

Otu24 Unclassified Actinomycetales

0.88

1

Otu29 Unclassified Sphingobacteriales

0.88

1

Otu31 Unclassified Cryomorphaceae

0.898

1

Otu41 Unclassified Microbacteriaceae

0.83

1

Otu42 Methylophilus

0.78

1

Otu45 Unclassified Cytophagaceae

0.88

1

Otu48 Unclassified Sphingomonadaceae 0.74

Aeromonas

97

Figure 2.1: Mean proportion of cumulative survival (solid-line) of lake sturgeon
from Transition (TR) and Control (CR) treatment group with 95% CI (dashedline). Survival analyses were based on log-rank tests indicating that survival of fish in the
TR treatment (0.527) was significantly lower compared to fish from the CR treatment
(0.987) (Chi square value, χ2 = 85.1, p-value < 0.01, df =1)

98

(a)

Length (mm)

50

45

Treatment
Group
Control
Transition

40

ns
35

30
14dpf

21dpf

35dpf

Days of active feeding

Weight per indiv. fish (g)

(b)0.5

0.4

0.3
Treatment
Group
Control

0.2

Transition

ns
ns

0.1

1

2

3

4

5

Week of active feeding

6

Figure 2.2. Morphometric data with error bar representing standard error mean (a)
total length (mm) by treatment group, at each sampling point; (b) average weight

99

Figure 2.2 (cont’d)
(g) of fish from each treatment group fed with control food vs fish transitioned to
Chironomid larvae. * indicate significant differences in treatment mean test based
on Welch-t test (* p-value < 0.05; ** p-value < 0.01; *** p-value < 0.001)

100

(a)

14dpf

21dpf

36dpf

Treatment CR

Phyla
Acidobacteria
Actinobacteria
Bacteroidetes
Chloroflexi
Firmicutes
Fusobacteria
Proteobacteria

Treatment TR

(b)

14dpf
Water

21dpf
Water

unclassified
Verrucomicrobia
other

36dpf
Water

phylum
Actinobacteria
Bacteroidetes
Firmicutes
Proteobacteria
unclassified
Verrucomicrobia
other

Figure 2.3 (a) Relative abundance (percentage) of six bacterial phyla (Acidobacteria,
Actinobacteria, Bacteroidetes, Firmicutes, Fusobacteria, Proteobacteria) found in gut
microbiota of lake sturgeon larvae for each treatment at different times. The
remaining taxa were assigned as Others. (b) Relative abundance (percentage) of six
bacterial phyla (Acidobacteria, Actinobacteria, Bacteroidetes, Firmicutes,
Proteobacteria, Verrucomicrobia) found in aquatic samples. The remaining taxa
were assigned as Others.

101

(a)

50
45

Inverse Simpson index, 1/D

40
35
30
25

Treatment
Control, CR

20

Transition, TR
15
10
5
0
14dpf

21dpf

36dpf

Days of active feeding

(b)

2000
1900
1800
1700
1600
1500
1400

Observed taxa

1300
1200
1100
1000

Treatment

900

Control

800

Transition

700
600
500
400
300
200
100
0
14dpf

21dpf

36dpf

Days of active feeding

Figure 2.4 (a) Inverse Simpson Diversity index and (b) number of observed OTUs/
Taxa richness for both feeding treatment (Control, CR and Transition, TR) at
different sampling time

102

(a)

(b)

0.4

Treatment
Control
Transition
Water

0.2

Stages
14dpf
0.0

21dpf
36dpf

PCoA3, Var explained=12.54%

PCoA2, Var explained=21.06%

0.6

Treatment

0.50

Control
Transition
Water
0.25

Stages
14dpf
21dpf
36dpf

0.00

−0.2

−0.2

0.0

0.2

−0.2

0.4

0.0

0.2

0.4

PCoA1, Var explained=23.91%

PCoA1, Var explained=23.91%

Figure 2.5 Visualization using Multivariate Principal Coordinates Analyses (PCoA) plots of variation in lake sturgeon gut and
water microbial community composition among collections made at different times and between food from different treatment

103

PCoA3, Var explained=12.54%

(c)

Treatment

0.50

Control
Transition
Water
0.25

Stages
14dpf
21dpf
36dpf

0.00

−0.2

0.0

0.2

0.4

0.6

PCoA2, Var explained=21.06%

Figure 2.5 (cont’d)
groups using Bray-Curtis distances. Samples were taken from replicates of different treatments at three developmental stages
(before transition at 14dpf, during transition week at 21dpf, after transition at 36dpf). (a) Plot based on Bray Curtis distance
Axis 1 and 2, (b) Plot based on Bray Curtis distance Axis 1 and 3, (c) Plot based on Bray Curtis distance Axis 2 and 3

104

(a)

(b)

Figure 2.6 Relative abundance of top 20 annotated functional groups identified
using KEGG Orthologs pathway categorized at molecular-level 2 associated with

105

(c)

Figure 2.6 (cont’d)
lake sturgeon larval microbiome from two diet treatments, Control (CR) and
Transition (TR) at each sampling point. (a) 14 dpf; (b) 21 dpf; (c) 36 dpf

106

(a)

(b)

Figure 2.7 Weighted Neighbor-Joining tree showing evolutionary relationship for
selected top match taxa with single outgroup species. Branch supports were
estimated using bootstrap analysis based on 100 pseudoreplicates for NJ and the

107

Figure 2.7 (cont’d)
percentage of bootstrap value were displayed next to branches in the inferred
phylogeny. (a) NJ tree for isolates from fish in the CR group; (b) NJ tree for
isolates from fish in the TR group.

108

LITERATURE CITED

109

LITERATURE CITED
Agh, N., Noori, F., Irani, A., & Makhdom, N. M. (2012). First feeding strategy for
hatchery produced Beluga sturgeon , Huso huso larvae. Iranian Journal of Fisheries
Sciences, 11: 713–723.
Agh, N., Noori, F., Irani, a., Van Stappen, G., & Sorgeloos, P. (2013). Fine tuning of
feeding practices for hatchery produced Persian sturgeon, Acipenser persicus and
Beluga sturgeon, Huso huso. Aquaculture Research, 44(3): 335–344.
Anderson, M. J. (2001a). A new method for non-parametric multivariate analysis of
variance, Australian Ecology, 26: 32-46
Anderson, M. J. (2006). Distance-based tests for homogeneity of multivariate
dispersions. Biometrics, 62: 245-253
Anderson, M. J. and Walsh, D. C. I. (2013). What null hypothesis are you testing?
PERMANOVA, ANOSIM, and the Mantel test in the face of heterogenous
dispersions. Ecology monographs, 83: 557-574
Bakke, A. M., Glover, C., & Krogdahl, Å. (2010). Feeding, digestion and absorption of
nutrients. In Grossell, M., Farrel, A. P., Brauner, J. (Eds.) The Multifunctional Gut of
Fish : Fish Physiology. Vol. 30. Academic Press, Elsevier, London: 165-212
Balcázar, J. L., De Blas, I., Ruiz-Zarzuela, I., Cunningham, D., Vendrell, D., & Múzquiz,
J. L. (2006). The role of probiotics in aquaculture. Veterinary Microbiology, 114(3–
4): 173–186.
Bar, N. (2014). Physiological and hormonal changes during prolonged starvation in fish.
Canadian Journal of Fisheries and Aquatic Sciences, 1458(71): 1–12.
Bauman, J.M. (2015). Investigations of aquaculture methodologies to enhance success of
Great Lakes Lake Sturgeon Streamside Facilities. Master Thesis, Department of
Fisheries & Wildlife, Michigan State University, East Lansing, Mich., USA.
Bauman, J. M., Woodward, B. M., Baker, E. A., Marsh, T. L., & Scribner, K. T. (2016).
Effects of Family, Feeding Frequency, and Alternate Food Type on Body Size and
Survival of Hatchery-Produced and Wild-Caught Lake Sturgeon Larvae. North
American Journal of Aquaculture, 78(2): 136–144.
Bjornsdottir, R., Johannsdottir, J., Coe, J., Smaradottir, H., Agustsson, T.,
Sigurgisladottir, S., & Gudmundsdottir, B. K. (2009). Survival and quality of halibut
larvae (Hippoglossus hippoglossus L.) in intensive farming: Possible impact of the
intestinal bacterial community. Aquaculture, 286(1–2): 53–63.

110

Bray, J. R. and Curtis, J. T. (1957) An ordination of the upland forest communities of
southern Wisconsin. Ecological monographs. 27(4): 325-49
Buddington, R. K., & Christofferson, J. P. (1985). Digestive and feeding characteristics
of the chondrosteans. Environmental Biology of Fishes, 14(1): 31–41.
Cain, K., & Swan, C. (2010). Barrier function and immunology. In Grossell, M., Farrel,
A. P., Brauner, J. (Eds.) The Multifunctional Gut of Fish : Fish Physiology. Vol. 30,
pp: 111-134. Academic Press, Elsevier, London
Callman, J. L., & Macy, J. M. (1984). The predominant anaerobe from the spiral intestine
of hatchery-raised sturgeon (Acipenser transmontanus), a new Bacteroides species.
Archive of Microbiology, 140, 57–65.
Czeskleba, D. G., Avelallemant, S., & Thuemler, T. F. (1985). Artifical spawning and
rearing of lake sturgeon, Acipenser fulvescens, in Wild Rose State Fish Hatchery,
Wisconsin, 1982-1983. In F. P. Binkowski & S. I. Doroshov (Eds.), North American
Sturgeons, pp: 80–85. Netherlands: Dr W. Junk Publishers
Das, P., Mandal, S. C., Bhagabati, S. K., Akhtar, M. S., & Singh, S. K. (2012). Important
Live Food Organisms and Their Role in Aquaculture. Frontiers in Aquaculture,
(March 2017): 69–86.
De Schryver, P., & Vadstein, O. (2014). Ecological theory as a foundation to control
pathogenic invasion in aquaculture. The ISME Journal: 1–9.
Deng, D.-F., Koshio, S., Yokoyama, S., Bai, S. C., Shao, Q., Cui, Y., & Hung, S. S. .
(2003). Effects of feeding rate on growth performance of white sturgeon (Acipenser
transmontanus) larvae. Aquaculture, 217(1–4): 589–598.
DiLauro, M. N., Krise, W. F., & Fynn-Aikins, K. (1998). Growth and Survival of Lake
Sturgeon Larvae Fed Formulated Diets. The Progressive Fish-Culturist, 60(4): 293–
296.
Gómez, G. D., & Balcázar, J. L. (2008). A review on the interactions between gut
microbiota and innate immunity of fish. FEMS Immunology and Medical
Microbiology, 52(2): 145–154.
Hamlin, H. J., Michaels, J. T., Beaulaton, C. M., & Main, K. L. (2006). Refining Feeding
Practices for Hatchery Production of Siberian Sturgeon , Acipenser baeri. Journal of
the World Aquaculture Society, 37(2): 224–230.
Hamre, K., Yufera, M., Ronnestad, I., Boglione, C., Conceicao, L. E. C., & Izquierdo, M.
(2013). Fish larval nutrition and feed formulation: Knowledge gaps and bottlenecks
for advances in larval rearing. Reviews in Aquaculture, 5:S26-S58

111

Holt, G. J. (2011). Larval Fish Nutrition. Wiley-Blackwell, Oxford, UK.
Hooper, L. V., Littman, D. R., & Macpherson, a. J. (2012). Interactions Between the
Microbiota and the Immune System. Science, 336(6086): 1268–1273.
Hoseinifar, S. H., Ringø, E., Shenavar Masouleh, A., & Esteban, M. Á. (2014). Probiotic,
prebiotic and synbiotic supplements in sturgeon aquaculture: a review. Reviews in
Aquaculture, (August),
Hoshino, T., Ishizaki, K., Sakamoto, T., Kumeta, H., Yumoto, I., Matsuyama, H., &
Ohgiya, S. (1997). Isolation of a Pseudomonas species from fish intestine that
produces a protease active at low temperature. Letters in Applied Microbiology,
25(1): 70–72.
Hung, S. S. O., & Deng, D. (2002). Sturgeon, Acipenser spp . In C. D. Webster & C. Lim
(Eds.), Nutrient Requirements and Feeding of Finfish for Aquaculture, pp. 344–357.
CABI Publishing.
Jobling, M. (2016). Fish nutrition research: past, present and future. Aquaculture
International, 24(3): 767–786.
Jukes, T. H. and Cantor, C. R. (1969). Evolution of protein molecules. In Munro, H. N.,
(Eds.), Mammalian Protein Metabolism, pp: 21-132, Academic Press, New York.
Kjørsvik, E., Galloway, T. F., Estevez, A., Sæle, Ø., & Moren, M. (2011). Effects of
Larval Nutrition on Development. In Holt, G. J. (Eds) Larval Fish Nutrition, pp: :
219–248,Wiley-Blackwell, Oxford, UK
Klassen, C. N., & Peake, S. J. (2007). The use of black fly larvae as a food source for
hatchery-reared lake sturgeon. North American Journal of Aquaculture, 69(3), 223–
228.
Klassen, C. N., & Peake, S. J. (2008). Effect of diet switch timing and food source on
survival and growth of lake sturgeon. Journal of Applied Ichthyology, 24(5), 527–
533.
Kohl, K. D., Amaya, J., Passement, C. A., Dearing, M. D., & Mccue, M. D. (2014).
Unique and shared responses of the gut microbiota to prolonged fasting: A
comparative study across five classes of vertebrate hosts. FEMS Microbiology
Ecology, 90(3)
Kolkovski, S., Koven, W., and Tandler, A. (1997). The mode of action of Artemia in
enhancing utilization of microdiet by gilthead seabream Sparus aurata larvae.
Aquaculture, 155: 193-205

112

Kozich, J. J., Westcott, S. L., Baxter, N. T., Highlander, S. K., & Schloss, P. D. (2013).
Development of a dual-index sequencing strategy and curation pipeline for
analyzing amplicon sequence data on the miseq illumina sequencing platform.
Applied and Environmental Microbiology, 79(17), 5112–5120.
Langdon, C., & Barrows, R. (2011). Microparticulate diets: technology, In Holt, G. J.
(Eds) Larval Fish Nutrition, Wiley-Blackwell, Oxford, UK 335–351.
Langille, M. G. I., Zaneveld, J., Caporaso, J. G., McDonald, D., Knights, D., Reyes, J. A.,
Clemente, J. C., Burkepile, D. E., Thurber, R. L. V., Knight, R., Beiko, R. G.,
Huttenhower, C. (2013). Predictive functional profiling of microbial communities
using 16S rRNA marker gene sequences. Nature Biotech, 31(9): 814–821.
Lazo, J. P., Darias, M. J., & Gisbert, E. (2011). Ontogeny of the Digestive Tract. In Holt,
G. J. (Eds) Larval Fish Nutrition, pp : 3–46, Wiley-Blackwell, Oxford, UK.
Léger, P., Bengston, D. a., Sorgeloos, P., Simpson, K. L., & Beck, A. D. (1987). The
nutritional value of Artemia: a review. In P. Sorgeloos, D. A. Bengston, W. Decleir,
& E. Jaspers (Eds.), Artemia Research and its Applications (Vol. 3, pp. 357–372).
Wetteren, Belgium: Universal Press.
Li, S., & Mathias, J. A. (1987). The Critical Period of High Mortality of larvae fish - a
discussion based on current research. Chinese Journal of Oceanology and
Limnology, 5(1): 80-96
Llewellyn, M. S., Boutin, S., Hoseinifar, S. H., & Derome, N. (2014). Teleost
microbiomes: The state of the art in their characterization, manipulation and
importance in aquaculture and fisheries. Frontiers in Microbiology, 5 (207):1-17.
Macpherson, A. J., & Harris, N. L. (2004). Interactions between commensal intestinal
bacteria and the immune system. Nature Reviews. Immunology, 4(6): 478–485.
M Hixson, S. (2014). Fish Nutrition and Current Issues in Aquaculture: The Balance in
Providing Safe and Nutritious Seafood, in an Environmentally Sustainable Manner.
Journal of Aquaculture Research & Development, 3(3): 21–23.
McCue, M. D. (2010). Starvation physiology: Reviewing the different strategies animals
use to survive a common challenge. Comparative Biochemistry and Physiology - A
Molecular and Integrative Physiology, 156(1): 1–18.
McCue, M. D. (2012). Comparative Physiology of Fasting, Starvation, and Food
Limitation. Springer New York.
Merrifield, D. L., Dimitroglou, A., Foey, A., Davies, S. J., Baker, R. T. M., Bøgwald, J.,
Castex, M., and Ringø, E. (2010). The current status and future focus of probiotic

113

and prebiotic applications for salmonids. Aquaculture, 302(1-2), 1–18.
Milligan-Myhre, K., Charette, J. R., Phennicie, R. T., Stephens, W. Z., Rawls, J. F.,
Guillemin, K., & Kim, C. H. (2011). Study of host-microbe interactions in zebrafish.
In Methods in cell biology, 105: 87-116
Mim, S. D., Lazur, A., Shelton, W. L., Gomelsky, B., & Chapman, F. (2002). Production
of sturgeon-South Regional Aquaculture Center.pdf. Southern Regional Aquaculture
Center, 7200: 1-8
Naylor, R. L., Hardy, R. W., Bureau, D. P., Chiu, A., Elliott, M., Farrell, A. P., Forster, I.,
Gatlin, D. M., Goldburgh, R. J., Huac, K., Nichols, P. D. (2009). Feeding
aquaculture in an era of finite resources. Proceedings of the National Academy of
Sciences of the United States of America, 106(36): 15103-15110
Legendre, P., and Legendre, L.F. (2012). Numerical ecology, 7:18
Ohs, C. L., Cassiano, E. J., & Rhodes, A. (2009). Choosing an Appropriate Live Feed for
Larviculture of Marine Fish. IFAS Extension University of Florida, 1–7.
Oie, G., Reitan, K. I., Evjemo, J. O., Stottrup, J., & Olsen, Y. (2011). Live Feeds. In Holt,
G. J. (Eds) Larval Fish Nutrition, pp: : 307–334. Wiley-Blackwell, Oxford, UK
Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O'Hara, R. B.,
Simpson, G. L., Solymos, P., Stevens, M. H. H., and Wagner, H. (2015). vegan:
Community Ecology Package. R package version 2.2-1.
Olson, M.H. (1996). Ontogenetic niche shifts in largemouth bass: variability and
consequences for first-year growth. Ecology, 77: 179-190.
Parks, D. H., Tyson, G. W., Hugenholtz, P., & Beiko, R. G. (2014). STAMP: statistical
analysis of taxonomic and functional profiles. Bioinformatics, 30(21): 3123–3124.
Rabe, J., & Brown, J. A. (2000). A pulse feeding strategy for rearing larval fish : an
experiment with yellowtail flounder, Aquaculture, 191(4): 289–302.
Ringø, E., Zhou, Z., Vecino, J. L. G., Wadsworth, S., Romero, J., Krogdahl, J., Olsen, R.
E., Dimitroglou, A., Foey, A., Davies, S., Owen, M., Lauzon, H. L., Martinsen, L.
L., Bossier, P., Sperstad, S., and Merrifield, D. L. (2016). Effect of dietary
components on the gut microbiota of aquatic animals. A never-ending story?
Aquaculture Nutrition, 22: 219-282.
Romero, J., & Navarrete, P. (2006). 16S rDNA-based analysis of dominant bacterial
populations associated with early life stages of coho salmon (Oncorhynchus
kisutch). Microbial Ecology, 51(4): 422–430.

114

Saitou, N. and Nei, M. (1987). The Neighbour-joining method: A new method for
reconstructing phylogenetic trees. Molecular Biology Evolution, 4(4): 406-425
Sharifian Fard, M., Pasmans, F., Adriaensen, C., Laing, G. Du, Janssens, G. P. J., &
Martel, A. (2014). Chironomidae bloodworms larvae as aquatic amphibian food.
Zoo Biology, 33(3): 221–227.
Skjermo, J., & Vadstein, O. (1999). Techniques for microbial control in the intensive
rearing of marine larvae. Aquaculture, 177(1–4): 333–343.
Tetlock, A., Yost, C. K., Stavrinides, J., & Manzon, R. G. (2012). Changes in the gut
microbiome of the sea lamprey during metamorphosis. Applied and Environmental
Microbiology, 78(21): 7638–44.
Therneau, T. M. and Grambsch P. M. (2000). Modeling Survival Data:
Extending the Cox Model. Springer, New York.
Ursell, L. K., Metcalf, J. L., Parfrey, L. W., and Knight, R. (2012). Defining the Human
Microbiome. Nutrition Reviews, 70(Suppl 1): S38–S44.
Vedrasco, A., Lobchenko, V., Pirtu, I., & Billard, R. (2002). The culture of live food for
sturgeon juveniles, a mini review of the Russian literature. International Review of
Hydrobiology, 87(5–6): 569–575.
Volkman, E. T., Pangle, K. L., Rajchel, D. A., & Sutton, T. M. (2004). Hatchery
Performance Attributes of Juvenile Lake Sturgeon Fed Two Natural Food Types.
North American Journal of Aquaculture, 66: 105–112.
Wickham, H. (2009). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag,
New York.
Wood, C. M., & Bucking, C. (2010). The role of feeding in salt and water balance. In
Grossell, M., Farrel, A. P., Brauner, J. (Eds.) The Multifunctional Gut of Fish : Fish
Physiology. Vol. 30, pp : 165-212. Academic Press, Elsevier, London
Wu, H. J., & Wu, E. (2012). The role of gut microbiota in immune homeostasis and
autoimmunity. Gut Microbes, 3(1): 4–14.
Xia, J. H., Lin, G., Fu, G. H., Wan, Z. Y., Lee, M., Wang, L., Liu, X. J., and Yue, G. H.
(2014). The intestinal microbiome of fish under starvation. BMC Genomics, 15(1):
266-276
Yufera, M. (2011). Feeding behavior in larval fish. In Holt, G. J. (Eds) Larval Fish
Nutrition, pp: 284–305, Wiley-Blackwell, Oxford, UK.

115

Zambonino Infante, J. L., & Cahu, C. L. (2010). Effect of nutrition on marine fish
development and quality. Recent Advances in Aquaculture Research: 103–124

116

CHAPTER 3: CHANGES IN LAKE STURGEON GUT MICROBIOMES IN
RESPONSE TO CHEMOTHERAPEUTANT TREATMENTS
ABSTRACT
Prophylactic treatments using antibiotic, drugs, or chemical compounds are widely
applied in aquaculture to prevent disease outbreaks. Simultaneously, the vital roles of gut
microbiota in maintaining host physiological processes and homeostatic regulation have
been recognized and a growing number of studies have reported the damaging impacts of
antibiotics to gut microbial community composition and function. Here, we characterized
the gut microbial composition of an ecologically and economically important fish
species, the lake sturgeon (Acipenser fulvescens), during early larval stages in response to
weekly prophylactic treatments using three different chemotherapeutants commonly used
in aquaculture (Chloramine-T, hydrogen peroxide, salt followed by hydrogen peroxide)
relative to a control treatment. Gut microbiome composition in treated vs non-treated
larval fish from two different hatchery-crosses and multiple groups developed from wild
eggs collected at two spawning sites were analyzed using massively parallel next
generation sequencing based on the V4 region of the 16S rRNA gene. Results showed
that members of the phylum Firmicutes (unclassified Clostridiales and
Clostridium_sensu_stricto) and Proteobacteria were the dominant gut microbiota of all
fish samples regardless of treatment. We observed large variation in the diversity of lake
sturgeon microbiota between larvae from hatchery and wild origin using Principal
Coordinate Analyses (PCoA). PERMANOVA analyses indicated a significant interaction
between families/group and treatment. Regression analyses suggested effects of
treatments were dependent on the fish origin. The influence of host genotype and the
resilience of gut microbiota to prophylactic treatments are discussed.

117

INTRODUCTION
Widespread use of chemicals, drugs, and antibiotics (chemotherapeutants) is a rising
concern in aquaculture. With recent expansion and rapid growth of the aquaculture
industry, prophylactic use of chemicals and antibiotics application has been reported in
the aquaculture as a means to control disease outbreaks (Cabello, 2006; Mortazavi,
2014). While short-term benefits are often realized, there is great potential for damaging
impacts of these practices, as large amounts of veterinary drugs and chemotherapeutants
are passed into the soil and aquatic environments. The emergence of antibiotic resistance
threatens fish, terrestrial animals and human beings (Cabello, 2006; Fraise, 2002; Samira
& Guichard, 2006). In fish, antibiotic resistance poses a more serious problem than for
terrestrial animals because water readily supports and spreads bacterial pathogens
(Romero, Ringø, & Merrifield, 2014).
Another important aspect of prophylactic treatment practices in aquaculture is the
potential impact of drugs and chemotherapeutants on the gut microbiomes. Common
treatment strategies include the use of chemotherapeutants to treat infected fish following
visual detection of disease or in response to high mortality events. Alternatively, weekly
prophylactic chemotherapeutant treatments are used to reduced stress and reduce the
probability of pathogen infection (Bowker et al. 2011). Chemotherapeutants and
antimicrobial compounds used in prophylactic treatments have been shown to be
effective at reducing or preventing mortalities caused by pathogens. However, some
compounds are indiscriminant in their effects and may eradicate symbiotic and
/commensal gut microbial communities as well (Romero et al., 2014). Downstream
effects of antibiotic or chemical treatments on microbiomes are likely to have important

118

consequences to fish hosts. However, these effects are currently under-studied.
In humans, microbiomes within individual hosts usually varies in composition
across anatomical sites, and taxonomic composition can vary over time in response to
factors such as diet, physical activities and medication intake (Cho & Blaser, 2012; Dave
et al., 2012; The Human Microbiome Project Consortium, 2012). Among these factors,
exposure to antibiotics can have profound effects on resident microbial communities
inside human guts (Dave et al., 2012; Langdon, Crook, & Dantas, 2016). Several studies
reported changes in density or gut microbiome composition, for instance in infants who
receive antibiotics (Palmers et al. 2007). Dethlefsen et al. (2008, 2011) documented
pervasive effects of orally administered antibiotic to adult gut microbiomes, associated
with decreases in taxa richness and evenness.
Antibiotic use can save human lives and the lives of economically important fish
species. However, there can be collateral damage to beneficial gut microbes from overutilization of antibiotics (Becattini, Taur, & Pamer, 2016; Langdon et al., 2016).
Antibiotic treatments can threaten indigenous gut microbiota by eliminating certain
bacterial taxa, potentially resulting in ecological drift or community alteration that favor
the increase in abundance of certain taxa (Costello et al., 2012; Manichanh et al., 2010;
Panda et al., 2014). Adverse effects related to antibiotic use include pathogen resistance,
suppression of the immune system, increased rates of allergies, autoimmunity, and other
immune-inflammatory conditions (see Langdon et al., 2016 for review). We are now just
beginning to understand the functions of microbial communities. Therefore it is important
to expand studies of the impact of antibiotics to the stability of gut microbiomes in other
ecologically and economically important vertebrates, including fishes.

119

Community ecology places emphasis on studies of patterns in diversity,
abundance, and composition of species along with processes underlying these patterns
(Vellend, 2010). One core theory in community ecology involves drift, in which
community composition changes in response to disturbance (Leibold et al., 2004).
Disturbance can be defined as a “single disruptive event or set of events that significantly
changes ecological community structure and function” (Christian, Whitaker, & Clay,
2015; Leibold et al., 2004). Ecological communities can respond in one of four ways to
disturbance: (i) communities experience no change in composition following the
disturbance (resistance); (ii) community composition changes but then returns to its
original state (resilience); (iii) community composition changes but the new microbial
constituents maintain the same function as the original community (functional
redundancy); or (iv) microbial community composition changes, some taxa are extirpated
and original community function is lost (perturbed) (Cho & Blaser, 2012; Christian et
al., 2015; Francino, 2016; Langdon et al., 2016)
Advances in sequencing technology have allowed researchers to advance studies
pertaining to gut microbiomes beyond traditional medical and veterinary applications
(Dave et al., 2012; Ghanbari, Kneifel, & Domig, 2015). Complex species interactions
across space and over time pose challenges to our understanding of the ecological
organization and evolutionary importance of animal-bacterial interactions. From an
ecological perspective, the microbial community within an individual host can be viewed
as a local community colonized from a regional species pool (Adair & Douglas, 2017).
Combining microbiomes studies with the conceptual framework of community ecology

120

that has developed over the years offers insight to illuminate complexities of hostmicrobe interactions (Christian, Whitaker, & Clay, 2015; Costello et al., 2012).
Few studies have documented changes in a fish-associated gut microbial
community in response to chemical or antibiotic exposure. The effect of orally
administered antimicrobial compounds on gut microbiome was reported in several
important aquaculture species including rainbow trout (Onchorynkuss mykiss) using
culture methods (Austin & Al-Zahrani, 1988) and molecular-based methods (Navarrete et
al. 2008); hybrid tilapia (Oreochromis niloticus X O.aureus) (He et al., 2010); and gibel
carp (Carassius auratus gibelito) (Liu et al., 2012) (for extensive review, see Ringø et al.,
2015). Collectively, these studies reported gut microbial communities are negatively
impacted by antimicrobial treatments. These studies, however, have focused mainly on
describing gut microbiome in fish at the juvenile stage, whereas fish at early life stages
are prone to pathogen infection (Vadstein et al., 2012), and thus may often be exposed to
antimicrobial compounds and chemotherapeutants. To evaluate the suitability of
prophylactic treatment of fish larvae without compromising fish normal function, more
studies are warranted pertaining to the influence of chemotherapeutants utilized in fish
culture on gut microbiota.
We empirically evaluated the response of lake sturgeon larval gut microbial
communities raised in a common rearing environment to perturbations associated with
the use of four prophylactic chemotherapeutants using 16S rRNA-based next generation
sequencing. We measured the changes in gut community composition of fish produced
from either wild origins or from gametes fertilized in the hatchery. We hypothesized that
prophylactically treated fishes would show decreased microbial taxonomic diversity and

121

different community composition relative to fish from the control treatment. We used
one-to-all comparison analyses (Segata et al. 2011) to identify biomarkers (taxa)
associated with each treatment in contrast to fish from the control group. Our findings
provide insight into the consequences of prophylactic treatments and host-microbe
interactions.

122

METHODOLOGY
Fish husbandry
The experiment was performed at the Black River Sturgeon Rearing Facility managed by
the Michigan Department of Natural Resources (MDNR) and Michigan State University
(MSU) in Onaway, MI, USA using approved Institutional Animal Care and Use
Committee (IACUC) protocols. This experiment was conducted from June 26 to July 30,
2013, following the primary spawning period for lake sturgeon in the Upper Black River,
Cheboygan County, Michigan.
Different sources of lake sturgeon progeny were utilized in this experiment. First,
hatchery-produced, full-sibling lake sturgeon larvae were produced from direct gamete
takes from two mated pairs of males and females collected during sturgeon spawning
season in the Upper Black River. The fish families were labeled as hatchery A (HA) and
hatchery D (HD). Gamete collection and eggs fertilizations followed procedures as
described by Crossman et al. (2011) and Bauman et al. (2015), respectively. Secondly,
wild, naturally produced larvae from multiple spawning events were raised from wild
naturally fertilized eggs that were collected from the stream substrate from two spawning
locations at the Upper Black River approximately three days post-fertilization. These
eggs were transported to the hatchery, incubated and separated by capture location
labeled as Site B and Site C, and fish cohort hatched from these spawns were denoted as
wild B (WB) and wild C (WC). Parentage was not determined but previous studies
indicated that fertilized eggs are likely from multiple pairs at each site.
Eggs were incubated and raised in similar hatchery rearing environments
separately based on their family using Aquatic Eco-systems (Pentair) J32 Mini-Egg

123

hatching jars. Water was taken directly from the river and filtered through 50 micron and
100 micron filters to remove sediment and biological material. Stream water was used
throughout the experiment. Eggs were treated daily with 500 ppm, 15 min immersion in
hydrogen peroxide as part of standard incubation procedure until 24 hours prior to hatch.
At hatch, free-embryos were transferred into 10L polycarbonate tanks (Aquatic Habitats)
consisting of sinking Bio-Balls (Pentair #CBB 1-S) that covered the bottom of each tank.
As endogenous yolk resources were depleted and larval fish emerged, Bio-Balls were
removed and fish were fed brine shrimp three times daily. Feeding rate was determined
following Deng et al. (2003). Ambient mean water temperature during the study was 22.7
o

C(range 20.0 to 26.3 oC). Mortalities were recorded daily, and all individuals were raised

in the same rearing conditions until the prophylactic treatment experiment began.
Experimental design
We used four treatment groups where fish in each group were exposed to weekly
prophylactic treatments using compounds that are commonly used in aquaculture
(collectively called chemotherapeutants). Treatments included: 1) 60 min, 15 ppm
Chloramine-T (CT) immersion, 2) 15 min, 60 ppm hydrogen peroxide (H2O2), 3) 3 parts
per thousand (ppt) salt, NaCl- immersion for 15 min followed 24 hr later by a 15 min, 60
ppm H2O2 bath (NaCl/H2O2), and 4) a control group (no chemical treatment, CTRL). The
experimental design associated with how the fish were quartered between treatments is
shown in Fig. 3.1.
At twelve days post feeding (dpf), 400 fish from each hatchery family or wild
group were transferred and randomly distributed into four 1.2 m diameter tanks which
were divided into eight partitions with 50 fish per partition. There were two replicate

124

groups of 50 fish in each treatment group in each tank (Fig. 3.1). The first prophylactic
exposure began at 14 dpf (after two days of tank acclimation). Treatments were
conducted at 5 weekly intervals and continued for 35 days (total 5 treatments).
Assignment of disinfectant exposure was based on a randomized block designed where
disinfectant exposure was randomly assigned to each partition within the tank (Fig. 3.1).
Fish mortalities were recorded and total survival was documented at the end of the
experiment. Comparisons were made of the effects of different prophylactic treatments
on diversity and compositional change in bacterial communities.
During immersion, all fish from each treatment group (including controls) were
transferred using a hand net to 10L polycarbonate tanks to expose fish to
chemotherapeutants (or river water for the control group). Following each treatment, fish
were briefly rinsed and immediately placed back into their assigned rearing tank and
compartment. All exposures were administered on the same day and were repeated once
per week except treatment 3 (salt), which included an additional treatment the following
day with H2O2. Fish in the control group were handled in the same manner as fish from
other treatment groups, however, similar to treatment 1 (CT), were held for 60 min in
their ‘treatment’ tank before being rinsed and returned to their rearing tanks. Sampling
for microbiota analysis took place following the end of the five-week treatment period.
From each partition, four fish were randomly collected and were euthanized with an
overdose of MS-222 (Sigma-Aldrich, St Loius, MO, USA). All fish were preserved in
80% ethanol and transported to MSU until dissections were performed.

125

Fish dissection, DNA isolation, PCR validation
The distal gut (spiral valve) of each sturgeon larvae was recovered from fish following
aseptic techniques. The distal gut was defined as the section that begins at the intestine
until the spiral valve. The spiral valve serves as the primary region of digestion and
absorption, and thus may provide an area of abundant nutrients where a microbial
community can flourish (Buddington & Christofferson, 1985; Callman & Macy, 1984).
Exterior surfaces were swabbed with 100% ethanol before dissections of the whole
digestive tract using sterile instruments. Dissections were performed with slight
modification as previously described by (Milligan-Myhre et al., 2011). The intact
alimentary tracts were cut from the body cavity, and the excised gut was immediately
transferred into filtered-sterilized 80% ethanol solution for DNA isolation. Until DNA
extraction was performed, all samples were stored in -20oC.
Each gut sample was first centrifuged for 15 min at 4oC to pellet bacteria before
DNA was extracted. The MoBio PowerSoil® DNA Isolation Kit (Carlsbad, CA, USA)
including a bead-beating step was used following protocols for low-biomass samples, as
suggested by the manufacturer. The integrity of each DNA sample was assessed based on
amplification of 1.4k bp of the 16S rRNA gene (amplicon based on 27F and 1389R
primers) followed by gel agarose electrophoresis. DNA concentrations were quantified
using a Microplate spectrophotometer (BioTek®, Winooski, VT, USA).
16S rRNA amplicon sequencing and sequence pipeline analyses
Gut microbiota from lake sturgeon larvae were surveyed using high-throughput
sequencing of the 16S rRNA gene. 152 DNA samples (including four positive controls,
and 20 technical replicates) that have been validated to contain sufficient bacterial DNA

126

(as shown by the presence of amplicon bands in electrophoresis) were submitted for
sequencing at Michigan State University Research Technology Support Facility, RTSF
(East Lansing, MI, USA). All sequencing procedures, including the construction of
Illumina sequencing library, emulsion PCR, and MiSeq paired-end sequencing v2
platforms of the V4 region (~250bp; primer 515F and 806R) followed standard Illumina
(San Diego, CA, USA) protocols. Michigan State Genomics RTSF
(https://rtsf.natsci.msu.edu/genomics/) provided standard Illumina quality control,
including base calling by Illumina Real Time Analysis v1.18.61, demultiplexing, adaptor
and barcode removal, and RTA conversion to FastQ format by Illumina Bcl2Fastq v1.8.4.
Details of the microbial sequence data analyses pipeline and computing workflow
were made following default steps. Briefly, paired-end sequence merging, quality
filtering, “denoising", chimera checking, and pre-cluster steps were conducted using an
open-source workflow based on methods implemented by program mothur v.1.36.1
(Kozich, Westcott, Baxter, Highlander, & Schloss, 2013). Then, sequence pipeline
analyses were performed in mothur v.1.38.1 to accomplish reference-based OTU
clustering (method = opticluster). Taxonomic assignment was performed using the
SILVA bacterial reference database by clustering sequences defined with 97% identity.
Any sequence singletons that were detected were removed prior to downstream analyses.
Rarefaction analyses were performed to evaluate the sampling coverage for each sample
based on the selected sequence depth. To minimize effects of under-sampling while
maintaining as broad dataset as possible, the final OTUs table was rarefied to a depth of
8000 sequences per sample. Nine DNA samples with low sequence depth were discarded
prior to downstream analyses.

127

Analyses of bacterial community profiles and ecological statistical analyses
3.1 Alpha diversity
All measures of microbial community diversity including Inverse Simpson (1/D)
diversity indices and OTUs richness of each sample were calculated from the sequence
data within program mothur. These indices were first analyzed using linear mixed-effects
regression (LMER) with the lme4 package in R (Bates et al. 2013). In contrast to a more
traditional approach to data aggregation and classical ANOVA analysis, LMER controls
for the variance associated with random factors without data aggregation (Baayen et al.
2008; Bolker et al. 2009). We started analyses with a null model that included diversity
indices as a dependent variable with family and wild groups selected as random factors.
We added chemical treatment as the predictor variable (fixed factor), and an interaction
between random and fixed effects to quantify whether model fit was improved. Model fit
was assessed using chi-square tests on the log-likelihood values to compare different
models (Winter, 2013). All analyses were carried out in the R program.
To test for significant differences in diversity indices of gut microbial
communities of fish across treatment groups (CT, H2O2, NaCl/ H2O2, CTRL) taking into
account the effect of the random (family) factor, we then used non-parametric KruskalWallis tests for the null hypothesis that diversity for each of fish samples did not differ as
a function exposure to a particular treatment. The test was followed by post-hoc TukeyKramer pairwise tests. P-values below 0.05 indicated significant differences in pairwise
comparisons. Identity and relative abundance and of all bacterial phyla in all fish gut and
water-associated microbial communities across sampling times were determined using
packages dplyr and reshape2 in program R.

128

3.2 Beta diversity
We used several packages implemented in program R to estimate comparative (beta [β])
diversity measures and ecological statistics at the OTU level. Briefly, we used the vegan
package (Oksanen, 2015) to generate Bray-Curtis (BC; Bray & Curtis, 1957) distance
characterizing differences in microbial community composition among samples. We used
the cmdscale function to perform Principle Coordinate analyses (PCoA) ordination of
community composition differences based on BC distance. The ggplot and ggplots2
packages (Wickham, 2009) were then used to create ordination plots to visually compare
gut community composition with aquatic community composition and as a function of
different treatments and among sampling periods based on the principal coordinates with
the first and second largest eigenvalues.
Multivariate hypothesis testing to quantify differences in community composition
among samples collected from different feeding treatments and across different families
were performed using the adonis function (Oksanen, 2015) in program R. Differences
between the centroids of sampling groups for treatments and family/samples origins were
based on Permutational Multivariate Analyses of Variance (PERMANOVA) using BC
resemblance matrices (Anderson, 2001; 2006). Under the null hypotheses, the centroids
of the groups (fish from either hatchery families and wild groups exposed to different
chemotherapeutant treatments) are equivalent for all groups under random allocation of
individual sample units to the groups (i.e., based on permutation). This test was employed
because of the non-parametric and skewed nature of microbial data.
To further analyze associations between family effects and chemical treatment
effects on fish gut microbiota, PCoA were performed separately on fish gut communities

129

within each family/group. Important PCoA axes denoted by larger eigenvalues in
comparison to average eigenvalue were selected, and linear regression models were
fitted, where each axis represents response variables given predictor variables of
chemotherapeutants. Under the null hypotheses, prophylactic treatments were not
expected to significantly affect fish gut community taxonomic composition within a
family. Axes here represent a linear combination of microbial community diversity and
composition present in individual fish GI tracts from the same family/group. Axes that
showed significant treatment effects were then analyzed using the lsmean function to
quantify effects of specified factors (or combinations of specified factors for interactions)
on the bacterial taxonomic composition and relative abundance in larval fish guts.
3.3 Differential abundance of OTUs and biomarker identification across treatments
To further determine the operational taxonomic units that most likely explained
differences in microbial community composition between treatment groups, we next
employed Linear Discriminant Analysis (LDA) effect size (LEfSe) methods (Segata et
al., 2011). In general, the LEfSe algorithm identifies genomic features (i.e., in the present
study – OTUs) that are differentially abundant in different experimental groups (families
and treatments), then ranks them based on that abundance differential. The larger the
difference in relative abundance between groups, the higher the importance of that OTU.
The algorithm first identifies features (OTUs) that were statistically different
among families/groups based on the nonparametric factorial Kruskal-Wallis (KW) sumrank test. Additional tests assessed consistency of differences using unpaired Wilcoxon
rank-sum tests. In the final step, LEfSe uses LDA to rank each differentially abundant
taxa in order of the difference in abundance based on an LDA Score (log-scale). Results

130

represent a scale indicating “importance” of an OTU in group differences in microbiota
composition (Segata et al., 2011).
To run LEfSe, the tabular file was generated from a shared file that contained no
singletons in program mothur v.1.39.5. The tabular file consisted of taxonomic relative
abundance in gut community samples from the four different families/groups assigned to
four chemotherapeutant treatments. Then, this tabular file was transferred into an online
bioinformatics toolkit developed by Huttenhower lab to perform LEfSe analyses
(https://huttenhower.sph.harvard.edu/galaxy/) and generate LDA figure.

131

RESULTS
Diversity of gut microbial community composition
A total of 144 samples were retained after quality filtering was performed in the sequence
pipeline analyses. Comparisons of gut microbial community composition of sturgeon fry
at the level of phyla indicated that three major phyla dominated more than 50% of total
community abundance across all fish samples (Firmicutes 16%, Proteobacteria 36.5%,
and Actinobacteria 15.1%). Phyla detected in a remainder of the gut community included
Acidobacteria, Bacteroidetes, and Verrucomicrobia along with other phyla that
collectively comprised 30% of communities.
The relative abundance of the most dominant phylum, Firmicutes was fairly
consistent across treatments for fish samples from all hatchery families and wild groups
(HA, HD and WB, WC, respectively). One exception was WB fish exposed to salt (mean
58%) and WC fish exposed to peroxide (mean 50%) that were relatively low compared to
other treatments (Fig. 3.2a). When comparing the abundance of Firmicutes across all four
families/groups, fish from hatchery family D (HD) had a lower percentage of Firmicutes
(mean range from 51-66%). Proteobacteria relative abundance was likewise relatively
uniform across treatments (13% - 28% of total abundance) with the exception of WB fish
that were treated with chloramine-T, CT (6%). Actinobacteria were present at 1% in fish
that were not exposed to any chemotherapeutant (control) but only in fish from HA and
WC groups.
At the genus level, Firmicutes were represented by two genera,
Clostridium_sensu_stricto & Unclassified genera from family Clostridiaceae. We found
that Clostridium_sensu_stricto were the most dominant genus (mean range: 30 – 51% of

132

the total community) for all fish of hatchery origin (except for HA fish exposed to
peroxide), whereas all fish of wild origin had Unclassified taxa from Clostridiaceae
family (mean range: 29 – 62%) as the most abundant genus across any treatment (Fig.
3.2b). Genera from phylum Proteobacteria including several unclassified taxa from
Alphaproteobacteria, unclassified taxa from Betaproteobacteria, unclassified taxa from
Enterobacteriaceae, unclassified taxa from Rhodobacteriaceae, and Deefgea all were
present at lower percentages of abundance with more amount of variation across fish
groups and treatments (Fig. 3.2b). The only genus in the phylum Actinobacteria that was
detected among dominant taxa was the genus Zhihengliuella, present in HA control fish
(mean 2.2%) and WC control fish (mean 1.4%).
To answer question, does treatment, or family or both affect alpha diversity
measures, we first generated three alternative models consisting of (i) both fixed
(prophylactic treatments) and random (family/group) predictor variables with interaction,
(ii) both fixed and random predictor variables without interaction, (iii) random predictor
variable only, to best explain variation in Inverse Simpson diversity indices and in the
number of observed taxa in fish gut microbial communities. For Inverse Simpson indices,
the model that consisted only of the random variable (family/group of origin) had the
lowest AIC value (AIC = 917.20, df =3) compared to the other two models that consisted
of fixed effects with and without an interaction (Table 3.1a). The log-likelihood test
indicated that no models evaluated were statistically different from one another.
The lowest AIC selected model for a number of observed taxa (OTU richness)
included both fixed and random variables (AIC= 1893.4, df =6). Comparisons of the
best-supported model with other models based on a log-likelihood test indicated that this

133

model was a significant predictor (χ2 = 17.71, df = 3, p < 0.001) (Table 3.1b). This
indicated a significant relationship between the number of microbial OTUs observed and
the family/group of fish samples.
Fig. 3.3 (a) and (b) shows the comparison of Inverse Simpson indices and number
of observed taxa for each chemical treatment and families/groups. Fish in the control
treatment (CTRL) had less diversity gut communities (both Inverse Simpson and
richness) with the exception of fish in family HD. However, Kruskal-Wallis tests for
Inverse Simpson for each hatchery family and wild group failed to reject the null
hypothesis that diversity does not differ significantly across treatments.
When the same statistical test was applied to fish from wild groups based on
observed numbers of taxa (OTU richness), we detected significant p-value therefore
rejecting the null hypothesis. Fish from one treatment group were characterized by taxa
richness that statistically differed from than other treatment in each wild group (WB, χ2 =
14.38, df = 3, p < 0.01 and WC, χ2 = 8.54, df = 3, p < 0.05). Post hoc tests (without
adjusted p-values) indicated that fish in the WB group that were exposed to salt treatment
were characterized by a significantly higher number of taxa when compared to WB fish
in the control treatment (salt-CTRL p < 0.01). Post-hoc comparisons indicated that no
differences among groups were detected in other treatments. For WC, we found that fish
exposed to peroxide had a significantly greater number of taxa relative to fish from the
control group (peroxide-CTRL p < 0.05), but not significantly different when compared
to the other two treatments.

134

Associations between gut microbial community composition among fish groups and
prophylactic treatments
Principal Coordinate Analyses (PCoA) ordination of differences in microbial taxonomic
composition and relative abundance of gut communities was performed to visualize
relationships between community composition associated with fish from different
families/groups and fish exposed to different chemotherapeutant treatments. Similarities
in community membership across samples of similar origin (either from the wild, or from
the hatchery production) were evident regardless of treatment groups as denoted by the
ordination pattern (Fig. 3.4). Ordination of sample communities associated with
treatments overlap considerably (see Fig. 3.4). Fish from eggs collected from the wild
(WB and WC) exhibited considerable variation in community composition relative to
variation among fish originating from hatchery crosses (HA and HD).
To test if there was any influence of chemical treatment on the gut communities,
PERMANOVA was performed. Statistical analyses of beta diversity across samples
showed that microbial communities present in the gut differed significantly based on the
significant interaction detected between families/groups (HA, HD, WB, WC) and
prophylactic treatments, as indicated by PERMANOVA test (Table 3.2: pseudo-F =
1.675, R2 = 0.787, p < 0.001). We reject the null hypothesis of no differences in
multivariate centroid location.
We next attempted to explain variation in community composition by
disentangling the treatment effect of different chemotherapeutants from the pervasive
effect of families/groups using linear regression and least square mean analyses. For fish
in family HA and wild group WB, their gut communities differed significantly for at least

135

one treatment (HA, p < 0.01; WB, p < 0.05) (See Table 3.3). For Family HA, the
interaction plot indicated that community membership of fish in the control group
differed from fish exposed to the other three chemotherapeutants (p < 0.01; Fig. 3.5a,
Table 3.3). Community membership of fish from WB associated with the control
(p<0.05) and salt treatments (p<0.01) differed significantly (Fig. 3.5b, Table2).
Identification of bacterial taxa influenced by chemotherapeutant treatments
Comparison of taxonomic abundance across all treatments and families/group (allagainst-all) in LEfSe did not yield any taxa that differed significantly in relative
abundance.
We conducted one-against-all comparisons using the control as a reference group
to compareo other treatments. Given the interaction between samples (family/group) and
prophylactic treatments influencing the gut community compositions as shown by
lsmeans analyses (for both family HA and origin WB), we used LEfSe to identify which
taxonomic groups that showed the largest differences in relative abundance when fish
from the same origin were exposed to treatments.
We first compared communities from fish from the control exposure group from
hatchery family, HA to fish exposed to other treatments at the genus level. We likewise
compared communities of fish from the control treatment within WB to other groups. We
found taxa associated two genera Methylocystaceae and Loktanella (both from phylum
Alphaproteobacteria) differed in abundance (LDA score higher than 2.0 or less than -2.0,
p < 0.05, data not shown) for comparison between fish communities in the control group
(CTRL) and peroxide (H2O2) fish (see Fig. 3.6a). Higher Loktanella were present in HA

136

fish samples from the CTRL treatment group (mean = 0.060) while Methylocystaceae
(mean=0.125) was significantly more abundance when exposed to H2O2.
LEfSe analyses performed with fish from the wild WB group detected three
differentially abundant taxa associated with genus Loktanella (Phylum
Alphaproteobacteria), Lysinibacillus (Phylum Firmicutes), and Unclassified Opitutae
(Phylum Verrucomicrobia). However, these three genera were present in high abundant
only in the gut of fish exposed to salt treatment with LDA score high than 2.0 (p < 0.05)
(see Fig. 3.6b). Average relative abundance of these genera that present in WB fish
individuals were Loktanella (mean = 0.027), Lysinibacillus (0.008), and Unclassified
Opitutae (0.027).

137

DISCUSSION
The presences of healthy, stable, and resilient gut microbial communities are
needed for normal functioning of the intestinal immune system and the general resistance
of the fish towards pathogens (Ringo et al. 2016) thus can contribute to sustainable and
successful production of cultured fish. However, for decades, animals raised in limited
space with high population densities for agricultural and aquaculture purposes are more
frequently exposed to chemicals and antibiotics which have been widely administered for
disease treatments, disease prevention, and to promote growth (Cabello, 2006).
Prophylactic treatments usually refer to antibiotic, drugs, or chemicals given to the
animals in order to prevent disease, as opposed to therapeutic treatments, which are
administered during pathogen outbreaks to treat the disease (Mortazavi, 2014).
In humans, several lines of evidence have confirmed that antibiotic administration
can result in gut microbiota dysbiosis (disturbance in composition and function; Francino,
2016) that could increase host susceptibility to disease infection. While more
comprehensive studies on adverse effects of antibiotic use to the gut microbiomes were
reported in human and other terrestrial animals used in agriculture (pig - Looft et al.,
2012; Schokker et al., 2015; ; chicken –Choi, Lee, & Sul, 2015; Schokker et al., 2017),
such data are scarce in fish, especially for important aquaculture species. Thus, the
present study was conducted to quantify the effects of prophylactic chemotherapeutants
on the gut microbial community composition of hatchery- and naturally-produced lake
sturgeon larvae. Using these findings, we could better inform industry professionals how
current practices impact fish health and performance in general.

138

To our knowledge, relatively few studies have been conducted addressing effects
of chemotherapeutants on gut microbial communities in general (not focusing on
pathogen), mostly in salmonids or tilapia (Austin & Al-Zahrani, 1988; Navarrete et al.
2010; He et al., 2010) and gibel carp (Carassius auratus gibelito) (Liu et al., 2012)
focusing on effects of antibiotic or drugs that were orally administered to confer internal
protection to fish hosts from harmful pathogenic bacteria. Using ecological theory, De
Schryver & Vadstein (2014) suggested that the primary compartment where pathogens
could be controlled is the water surrounding animals. We conducted studies focusing on
the impact of prophylactic treatments administered in the water on fish microbiota.
Fish and microbes live in close proximity within a shared aquatic environment.
Microbes respond rapidly to the changes in their immediate aquatic environment and
these changes could be subtle and manifested as activation or inactivation of certain
metabolic pathways. These changes could also be seen in changes in community
composition and functionality (Bentzon-Tilia, Sonnenschein, & Gram, 2016). However,
we were surprised to see that prophylactic treatment administered in our study had little
influence on the community composition of lake sturgeon gut microbiomes when
compared to gut communities of fish from the control group.
We detected three major phyla Firmicutes, Proteobacteria, and Actinobacteria,
that dominated the gut community (Figure 3.2a). Results were consistent with findings
from another study conducted to evaluate the effects of orally administered antibiotics to
gibel carp (Liu et al., 2012). In contrast to the previous study in our facility (Bauman,
2015 unpublished data), lake sturgeon eggs originated from three hatchery-produced
families that were prophylactically treated during incubation had an altered microbial

139

community on the egg surface. Analyses in our study indicated that the microbial
communities of fish across family/groups exposed to prophylactic treatments were
influenced to a larger extent by family/group (Fig. 3.2 and Fig. 3.3) than by
chemotherapeutant treatments. It is possible that the effect of treatment was not evident
due to the short treatment duration (15-60 min bath immersion) and periodicity of
treatments. Exposure to chemotherapeutant for such a brief period may not have been
strong enough to cause large-scale changes in gut community composition. In addition,
fish were returned into their tank partition aftertreatments, possibly allowing rapid
recolonization of gut microbiota from the surrounding water.
Major taxa that were detected from phyla Proteobacteria (such as
Enterobacteriaceae, Rhodobacteriaceae) are Gram-negative bacteria. Many clinical
studies have shown that Gram-negative bacteria are resistant to commercially available
antibiotics partly due to their thick cell wall structure compared to Gram-positive bacteria
(Slama, 2008; Vasoo, Barreto, & Tosh, 2015). Enterobacteriaceae include a group of
bacteria known as Extended Spectrum Beta-Lactamase (ESBL) Enterobacteriaceae that
confer resistance to antibiotics via production of the β-lactamase enzyme which can
inactivate certain β-lactam antibiotics (Jacoby & Munoz-Price, 2005).
Another major phylum, Firmicutes that were detected in fish guts across all
families and treatments was primarily represented by Unclassified Clostridiaceae1 and
taxa Clostridium sensu stricto. Although Clostridia are Gram-positive, these bacteria
have been identified as part of commensal gut microbiota that plays major roles in
maintenance the gut homeostasis. Several features possessed by Clostridium spp. could
explain why this taxon can thrive in the gut and can likewise be resistant to prophylactic

140

treatments administered in our study. In humans, Clostridium spp. are involved in
defenses inside intestinal microecosystem along with gut-associated lymphoid tissue
(GALT), and confer resistance against pathogen infections. This taxon is thought to have
immunological tolerance (Lopetuso et al., 2013). In addition, cultured Clostridium spp.
exhibit the ability to form endospores, which offers this bacteria with ecological
advantages for survival under adverse conditions (Gupta & Gao, 2009; Lopetuso et al.,
2013)
The influence of family/groups on gut microbial community composition was
most notable when comparisons were made based on alpha diversity indices (Fig. 3.3)
and visualization of beta diversity among all samples (Fig 3.4). All chemotherapeutants
used in our study are commonly used for a treatment of external pathogens rather than
orally administered to fish. Chloramine-T and peroxide are chosen to control and
eliminate infection associated with flavobacteriosis (Bowker et al. 2011). PERMANOVA
and least square mean revealed that chemotherapeutant treatments employed in our study
have only a minor effect on intestinal gut microbiome in fish, though effect varied among
fish with different genotype background associated with families/their sampling origin.
Recently, a study by Navarrete et al., (2012) assessed the relative contributions of a host
(genetics) and diet in shaping the gut microbiomes of rainbow trout. Full-sib fish from
four non-related families were fed two diet regimes in comparison to control group.
Results showed that some relative abundance of some bacterial groups differed among
trout families, indicating that the host genotypes may influence gut microbiota
composition. In addition, the authors reported that the effect of diet on microbiota
composition was dependent on the trout family. Studies on other organisms such as

141

chickens also showed that under a common diet and husbandry practices, gut microbiota
composition differed between two lines (High Weight, HW and Low Weight, LW) of
chicken with different quantitative genotypes indicated a prominent influence of host
genotype on community composition (Zhao et al., 2013).
Fish from artificially produced family HA experienced greater community
alteration compared than fish produced from wild eggs (Fig. 3.5a). Gut microbial
communities in wild fish possibly exhibited greater resilience to treatments and
maintained their gut compositional similarity. In contrast for fish hatchery, they
originated from eggs that have been artificially produced in enclosed facilities; therefore,
they had no contact with their respective natural habitat like the wild eggs, except their
egg surfaces reflect aquatic communities where their parents spawned. This could also
suggest that domestication selection in terms of hatchery gut community establishment
occurs in fish that were produced in the hatchery, affecting the community structure of
their gut microbiome. Findings from Blekhman et al. (2015) indicate that human gut
microbial variation are driven by host genetic variation involving genes that have been
previously associated with microbiome-related complex diseases. They also showed that
host genomic regions associated with microbiomes have high levels of genetic
differentiation among human populations, suggesting that host-genomic adaptation to
environment-specific microbiomes. This could be possibly true with fish as well where
variation in gut microbiome attributed to their genetic background and populations
differences.
In a companion study, Bauman (2015) quantified proportions of fish surviving
each of four prophylactic treatments. Fish exposed to salt followed by peroxide 24 hr

142

later were shown to have higher survival than other groups (Bauman et al. 2015).
Mortalities were observed throughout the course of the experiment. It is possible that gut
microbial communities of dead fish may have differed from fish surviving at the end of
experiment following the full course of the chemical treatment. However, since the dead
fish were not preserved and included in the studies, we have no empirical way to evaluate
this effect. Navarrete et al., (2008) reported that gut microbiomes of salmonids exposed
to Oxytetracycline (OTC) that were orally receiving antibiotic were characterized by less
diversity and were only composed of Aeromonas, clustering with A. sobria and A.
salmonicida. A. salmonicida has been known as a pathogenic bacteria that could result in
huge mortalities during the disease outbreak.
In the LEfSe analyses, only five out of thousands microbial taxa detected that
were differentially abundant after fish were exposed to chemical treatments. Those taxa,
however, are not among the dominant taxa. Thus it is unclear how treatment differentially
affected the relative abundance of these taxa. Results could provide indicate that the gut
microbiota were either resistant or exhibited resilience in community composition, where
treatment-based changes were short-lived and communities rapidly returned to their
original state (Christian et al., 2015). The communities could also have had different
compositional taxonomy, yet were still able to maintain function (functional redundancy).
Another study by Navarrete et al (2010) focused on determining effects of dietary
inclusion of Thymus vulgaris essential oil (TVEO) on microbiota composition, compared
with a control diet without TVEO over 5 week period. Their study indicated high
similarities between gut microbiota in treated and non-treated fish, and TVEO induced
negligible changes in gut microbiota profiles. Essential oils include volatile liquid

143

fractions produced by plants that contain the substances usually responsible for defenses
against pathogens and pests due to their antibacterial, antiviral, antifungal, and
insecticidal activities (Romero, Feijoó, & Navarrete, 2012). We conclude that gut
microbiota composition was persistent and stable throughout the trial, producing
temporally consistent molecular profiles. We documented the same outcome. The effect
of chemical treatment with regards to LEfSe result (Fig. 3.6a, 3.6b), it is possible that the
mode of action of each chemotherapeutant may explain the detection of biomarker taxa
by LEfSe. Salt treatment resulted in “shedding” the mucosal layers, and when followed
by peroxide, this prophylactic treatment could potentially remove certain bacterial taxa
underneath the mucosal layer.
Overall, our results suggest that treatments during larval stages do not result in
large changes in the composition of the intestinal microbiota, at least during the short
observation and experimental duration. It is possible that the effects of treatment on gut
communities were delayed until fish reach a later stage. Treatment was utilized to control
disease outbreaks in the hatchery, hence a prerequisite for developing a strategy for
microbial control, is knowledge of aquatic microflora associated with fish larvae, and
how interactions between larvae and microflora occur.
Our study serves as a baseline providing information on the indirect effects of
chemotherapeutant intervention that could either positively or negatively affect the
normal gut microbiota or facilitate the proliferation of opportunistic pathogens. Further
studies should be carried out comparing the effects of treatments before and after
chemotherapeutant are applied, focusing on the prevalence of potentially pathogenic
bacteria. Prophylactic effects of chemotherapeutants could also be performed coupled

144

with bacterial additives (e.g., probiotics) or challenges to document beneficial or
adversarial effects of these treatments on the gut microbiome composition.

145

APPENDIX

146

APPENDIX
Table 3.1 Models generated with both fixed and random variables to explain variation among samples of all treatments and
families/group (a) for Inverse Simpson diversity indices; (b) for a number of observed taxa (OTU richness).
The model with the lowest AIC scores was shown in bold.
Table 3.1 (a)
Residual

AIC

BIC

Log- likelihood

Residual
Deviance

df

χ2

p-value

Yinv ~ (1|Family)

3

917.20

926.11

-455.60

911.20

Yinv ~ Trt + (1|Family)

6

920.88

938.70

-454.44

908.88

2.315

0.510

Yinv ~ Trt + (1|Family) +

7

922.88

943.67

-454.44

908.88

0.000

1.000

Residual

AIC

BIC

Log- likelihood

Residual
Deviance

(1 | Trt :Family)
Table 3.1 (b)
df

χ2

p-value

Ysobs ~ (1|Family)

3

1905.1

1914.0

-949.55

1899.1

Y sobs~ Trt + (1|Family)

6

1893.4

1911.2

-940.70

1881.4

17.711 0.0005***

Ysobs ~ Trt + (1|Family) +

7

1895.4

1916.2

-940.70

1916.2

0.000

(1 | Trt :Family)

147

1.000

Table 3.2 PERMANOVA showing variability among fish gut microbiota across all samples. Results revealed a significant interaction
of both treatments (D) and family effect (F) influencing gut microbial communities composition for at least one samples across
treatments and families/group (PERMANOVA test pseudo-F = 1.675, R2 = 0.093, p < 0.001; permutation=1000).
Df
Treatment (T)
Family (F)
TXF
Residuals
TOTALS

3
3
9
128
143

Sum
Sq
0.870
3.097
3.046
25.866
32.880

Mean
Sq
0.290
1.032
0.338
0.202

F-model

R2

Pr (>F)

1.435
5.109
1.675

0.026
0.094
0.093
0.787
1.000

0.098
P<0.001***
P<0.001***

Table 3.3 Goodness of fit from linear regression model (R2), p-value, and Least-square means analyses performed on significantly
important PCoA axes calculated separately for each family/group. (a) PCoA Axis 2 associated with second largest eigenvalue for fish
from family HA show that microbial community composition in control group significantly differed than other treatments; (b) PCoA
Axis 6 associated with sixth largest eigenvalue for fish from group WB show that microbial community composition in both control
and salt fish significantly influence by each respective treatment

Linear regression model :
Pco Axes ~ Drug treatments
Important
axes
Family HA
Pco Ax.2
Family WB
Pco Ax.6

Signif. Treatment

p-value

Control (No drug)
Control (No drug)
Salt

0.008
0.012
0.001

Least-square means
R2 value
0.138
0.2425
0.2425

Control

Chloramine-T

Peroxide

Salt

-0.153±0.07

0.117±0.07

0.034±0.07

0.002±0.07

0.036±0.03

-0.081±0.03

-0.023±0.03

0.077±0.03

148

Figure 3.1 Schematic design of the larval chemotherapeutant study. Each 1.2m
diameter of tank held 400 fish from hatchery and wild naturally produced fish, which
were divided into eight equal sized partitions (50 fish per partition). There was four tanks.
Each partition was randomly assigned to one of four weekly treatment types, each with
two replicates. Chemotherapeutant treatments included: 1) 60 min, 15 ppm CT bath, 2)
15 min, 60 ppm H2O2, 3) 3 parts per thousand (ppt) NaCl- bath for 15 min followed 24 hr
later by a 15 min, 60 ppm H2O2 bath labeled as NaCl/ H2O2, and 4) a control (no
chemical treatment) labeled as CTRL. Arrows indicate directions of water flow. The
figure was originally from Bauman 2015 and is used here with permission.

149

(a)

Figure 3.2 Taxonomic composition of bacterial communities identified from the lake
sturgeon larval GI tracts (a) at the phyla level and (b) at the genera level.

150

(b)

Figure 3.2 (cont’d)
(a) Relative abundance (percentage) of dominant bacterial phyla found in the gut
microbiota of lake sturgeon larvae separated based on sample family/group to display
variation in communities across prophylactic treatments. Three predominant phyla were
present in gut microbial communities (Firmicutes, Proteobacteria, Actinobacteria). The
other phyla were characterized as Others. (b) Relative abundance (percentage) of
dominant bacterial taxa found in fish gut samples, separated by family/group and
treatment. Among the most abundant taxa included Unclassified Alphaproteobacteria,
Unclassified Betaproteobacteria, Unclassified Clostridiaceae_1,
Clostridium_sensu_stricto, and Unclassified Enterobacteriaceae.

151

(a)

Hatchery

Inverse Simpson index, 1/D

15

Wild

Treatment
Group

10

control
chloramineT
peroxide
salt

5

0

HA

HD

WB

WC

Family/Sample Origin
(b)

Hatchery

Wild

Number of observed taxa

500

400

Treatment
Group

300

control
chloramineT
peroxide
salt

200

100

0

HA

HD

WB

WC

Family/Sample Origin
Figure 3.3 Estimates of alpha diversity (a) Inverse Simpson index; (b) number of
observed taxa (OTU richness) for lake sturgeon gut microbial communities from all
samples of treatments and families/groups. Each bar indicates mean with S.E. for
each treatment from each family/group.

152

Figure 3.3 (cont’d)
Kruskal-Wallis followed by post-hoc test indicate that (i) none of families/treatments
differed for Inverse Simpson indices; (ii) for WB, fish that were exposed to salt treatment
have a significantly higher richness than fish from control group (salt-CTRL p < 0.01);
for WC, fish exposed to peroxide have a significantly higher richness to control group
(peroxide-CTRL p < 0.05).

153

PCoA2, Var explained=16.53%

0.50

Family/Sampling origin

HA
HD

0.25

WB
WC
0.00

Treatment

control
chloramineT
0.25

peroxide
salt
0.50

0.25

0.00

0.25

PCoA1, Var explained=30.69%

Figure 3.4 Visual representation of differences in the gut microbiota of lake
sturgeon larvae (Bray-Curtis distances). Bray-Curtis dissimilarity distances between
communities from fish exposed to four chemotherapeutant across families/group
were visualized by Principal Coordinates Analysis (PCoA) plots. Points represent
each gut microbial community from each individual fish samples from treatments and
families/groups ordinated in multivariate dimensional space using the first and second
axis corresponding to the first and second largest eigenvalues.

154

(a)

(b)

Figure 3.5 Interaction plot of marginal (least-square, LS) means for significant PCoA axes detected from hatchery family (HA)
and wild group B (WB). Axes represent variation in taxonomic diversity and relative abundance for a given family/group. (a)
A significant difference exists between microbial communities composition of HA fish in control group compared to fish that
received other prophylactic treatments; (b) Significant difference exists in the gut microbiota of fish from WB associated with
different prophylactic treatments. See details in Table 3.3.

155

(a)

(b)

Figure 3.6 LEfSe-detected taxa from one-to-all comparison (control to all other
three chemotherapeutant). (a) Thresholds on the logarithmic LDA scores for
discriminative features (taxa) in fish from family HA show that only two taxa were
differentially abundant in a comparison between CTRL and hydrogen peroxide treatment.
(b) Thresholds on the logarithmic LDA scores for discriminative features (taxa) in fish
from group WB detected that salt contain three taxa that were differentially abundant
compared to control group

156

LITERATURE CITED

157

LITERATURE CITED
Adair, K. L., & Douglas, A. E. (2017). Making a microbiome: the many determinants of
host-associated microbial community composition. Current Opinion in
Microbiology, 35: 23–29.
Anderson, M. J. (2001a). A new method for non-parametric multivariate analysis of
variance, Australian Ecology, 26: 32-46
Anderson, M. J. (2006). Distance-based tests for homogeneity of multivariate
dispersions. Biometrics, 62: 245-253
Anderson, M. J. and Walsh, D. C. I. (2013). What null hypothesis are you testing?
PERMANOVA, ANOSIM, and the Mantel test in the face of heterogenous
dispersions. Ecology monographs, 83: 557-574
Austin, B., & Al-Zahrani, a. M. J. (1988). The effect of antimicrobial compounds on the
gastrointestinal microflora of rainbow trout, Salmo gairdneri Richardson. Journal of
Fish Biology, 33(1): 1–14.
Baayen, R. H., Davidson, D. J., & Bates, D. M. (2008). Journal of Memory and
Language, 59: 390-412
Bates, D., Maechler, M., Bolker, B., Walker, S., Haubo, R., Christensen, B., Singmann,
H., Dai, B., Grothhendieck, G., & Green, P. (2013). lme4: Linear Mixed-Effects
Models using 'Eigen' and S4 Package. R package verison 1.1-13.
Bauman, J.M. (2015). Investigations of aquaculture methodologies to enhance success of
Great Lakes Lake Sturgeon Streamside Facilities. Master Thesis, Department of
Fisheries & Wildlife, Michigan State University, East Lansing, Mich., USA.
Bauman, J. M., Woodward, B. M., Baker, E. A., Marsh, T. L., & Scribner, K. T. (2016).
Effects of Family, Feeding Frequency, and Alternate Food Type on Body Size and
Survival of Hatchery-Produced and Wild-Caught Lake Sturgeon Larvae. North
American Journal of Aquaculture, 78(2): 136–144.
Becattini, S., Taur, Y., & Pamer, E. G. (2016). Antibiotic-Induced Changes in the
Intestinal Microbiota and Disease. Trends in Molecular Medicine, 22(6): 458–478.
Bentzon-Tilia, M., Sonnenschein, E. C., & Gram, L. (2016). Monitoring and managing
microbes in aquaculture – Towards a sustainable industry. Microbial Biotechnology,
9(5): 576–584.

158

Blekhman, R., Goodrich, J. K., Huang, K., Sun, Q., Bukowski, R., Bell, J. T., Spector, T.
D., Keinan, A., Ley, R. E., Gevers, D., & Clark, A. G. (2015). Host genetic variation
impacts microbiome composition across human body sites. Genome Biology, 16(1),
191.
Bolker, B. M., Brooks, M. E., Clark, C. J., Geange, S. W., Poulsen J. R., Stevens, M. H.
H., & White, S. S. S. (2009). Generalized linear mixed models: a practical guide for
ecology and evolution. Trends in Ecology and Evolution, 24: 127-135
Bowker, J., M. Gaikowski, A. Goodman, L. Sprague, and M. Bowman. (2011). Guide to
using drugs, biologics, and other chemicals in aquaculture. U. S. Fish and Wildlife
Service Aquatic Animal Drug Approval Partnership Program: 1-62.
Bray, J. R. and Curtis, J. T. (1957) An ordination of the upland forest communities of
southern Wisconsin. Ecological monographs. 27(4): 325-49
Buddington, R. K., & Christofferson, J. P. (1985). Digestive and feeding characteristics
of the chondrosteans. Environmental Biology of Fishes, 14(1): 31–41
Cabello, F. C. (2006). Heavy use of prophylactic antibiotics in aquaculture: A growing
problem for human and animal health and for the environment. Environmental
Microbiology, 8(7): 1137–1144.
Callman, J. L., & Macy, J. M. (1984). The predominant anaerobe from the spiral intestine
of hatchery-raised sturgeon (Acipenser transmontanus), a new Bacteroides species.
Archive of Microbiology, 140: 57–65.
Cho, I., & Blaser, M. J. (2012). The human microbiome: at the interface of health and
disease. Nature Reviews Genetics, 13(4): 260–270.
Choi, K. Y., Lee, T. K., & Sul, W. J. (2015). Metagenomic Analysis of Chicken Gut
Microbiota for Improving Metabolism and Health of Chickens — A Review. AsianAustralasian Journal of Animal Sciences, 28(9): 1217–1225.
Christian, N., Whitaker, B. K., & Clay, K. (2015). Microbiomes: unifying animal and
plant systems through the lens of community ecology theory. Frontiers in
Microbiology, 6(869): 1-19
Costello, E. K., Stagaman, K., Dethlefsen, L., Bohannan, B. J. M., & Relman, D. a.
(2012). The Application of Ecological Theory Toward an Understanding of the
Human Microbiome. Science, 336(6086): 1255–1262.
Dave, M., Higgins, P. D., Middha, S., & Rioux, K. P. (2012). The human gut
microbiome: Current knowledge, challenges, and future directions. Translational
Research, 160(4): 246–257.

159

De Schryver, P., & Vadstein, O. (2014). Ecological theory as a foundation to control
pathogenic invasion in aquaculture. The ISME Journal: 1–9.
Deng, D.-F., Koshio, S., Yokoyama, S., Bai, S. C., Shao, Q., Cui, Y., & Hung, S. S. .
(2003). Effects of feeding rate on growth performance of white sturgeon (Acipenser
transmontanus) larvae. Aquaculture, 217(1–4): 589–598.
Dethlefsen, L., Huse, S., Sogin, M.L., & Relman, D.A. (2008). The pervasive effects of
an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA
sequencing. PLoS Biology, 6:e280.
Dethlefsen, L. & Relman, D.A. (2011). Incomplete recovery and individualized responses
of the human distal gut microbiota to repeated antibiotic perturbation. Proceeding of
National Academy of Science USA, 108(Suppl 1):4554–4561.
Fraise, A. P. (2002). Biocide abuse and antimicrobial resistance--a cause for concern?
The Journal of Antimicrobial Chemotherapy, 49(1): 11–12.
Francino, M. P. (2016). Antibiotics and the human gut microbiome: Dysbioses and
accumulation of resistances. Frontiers in Microbiology, 6(1543): 1–11.
Ghanbari, M., Kneifel, W., & Domig, K. J. (2015). A new view of the fish gut
microbiome: Advances from next-generation sequencing. Aquaculture, 448: 464–
475.
Gupta, R. S., & Gao, B. (2009). Phylogenomic analyses of clostridia and identification of
novel protein signatures that are specific to the genus Clostridiumsensu stricto
(cluster I). International Journal of Systematic and Evolutionary Microbiology,
59(2): 285–294.
He, S., Zhou, Z., Meng, K., Zhao, H., Yao, B., Ringø, E. & Yoon, I. (2010). Effects of
dietary antibiotic growth promoter and Saccharomyces cerevisiae fermentation
product on produc- tion, intestinal bacterial community, and nonspecific immunity
of hybrid tilapia (Oreochromis niloticus female x Oreochromis aureus male).
Journal of Animal Science, 89: 84–92.
The Human Microbiome Project Consortium. (2012). Structure, function and diversity of
the healthy human microbiome. Nature, 486(7402), 207–214.
Jacoby, G. A., & Munoz-Price, L. S. (2005). The New β-Lactamases. New England
Journal of Medicine, 352(4): 380–391.
Kozich, J. J., Westcott, S. L., Baxter, N. T., Highlander, S. K., & Schloss, P. D. (2013).
Development of a dual-index sequencing strategy and curation pipeline for
analyzing amplicon sequence data on the miseq illumina sequencing platform.

160

Applied and Environmental Microbiology, 79(17): 5112–5120.
Langdon, A., Crook, N., & Dantas, G. (2016). The effects of antibiotics on the
microbiome throughout development and alternative approaches for therapeutic
modulation. Genome Medicine, 8(39):1-16.
Leibold, M. a., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J. M., Hoopes, M. F.,
Holt, R. D., Shurin, J. B., Law, R., Tilman, D., Loreau, M., & Gonzalez, a. (2004).
The metacommunity concept: A framework for multi-scale community ecology.
Ecology Letters, 7(7): 601–613.
Liu, Y., Zhou, Z., Wu, N., Tao, Y., Xu, L., Cao, Y., Zhang, Y., & Yao, B. (2012). Gibel
carp Carassius auratus gut microbiota after oral administration of
trimethoprim/sulfamethoxazole. Diseases of Aquatic Organisms, 99(3): 207–213.
Looft, T., Johnson, T. A., Allen, H. K., Bayles, D. O., Alt, D. P., Stedtfeld, R. D., Woo, J.
S., Stedtfeld, T. M., Chai, B., Cole, J. R., Hashsham, S. A., Tiedje, J. M., & Stanton,
T. B. (2012). In-feed antibiotic effects on the swine intestinal microbiome.
Proceedings of the National Academy of Sciences of the United States of America,
109(5): 1691–1696.
Lopetuso, L. R., Scaldaferri, F., Petito, V., & Gasbarrini, A. (2013). Commensal
Clostridia: leading players in the maintenance of gut homeostasis. Gut Pathogens,
5(23): 1-8
Manichanh, C., Reeder, J., Gibert, P., Varela, E., Llopis, M., Antolin, M., Guigo, R.,
Knight, R., & Guarner, F. (2010). Reshaping the gut microbiome with bacterial
transplantation and antibiotic intake. Genome Research, 20(10): 1411–1419.
Milligan-Myhre, K., Charette, J. R., Phennicie, R. T., Stephens, W. Z., Rawls, J. F.,
Guillemin, K., & Kim, C. H. (2011). Study of host-microbe interactions in zebrafish.
In Methods in cell biology, 105: 87-116
Mortazavi, A. L. (2014). Poppin' the prophylactics: An analysis of antibiotics in
aquaculture. The Journal of Global Health, 4(1): 23-27.
Navarrete, P., Magne, F., Araneda, C., Fuentes, P., Barros, L., Opazo, R., … Romero, J.
(2012). PCR-TTGE analysis of 16S rRNA from rainbow trout (oncorhynchus
mykiss) gut microbiota reveals host-specific communities of active bacteria. PLoS
ONE, 7(2): 1–10.
Navarrete, P., Mardones, P., Opazo, R., Espejo, R., & Romero, J. (2008). Oxytetracycline
treatment reduces bacterial diversity of intestinal microbiota of Atlantic salmon.
Journal of Aquatic Animal Health, 20(3): 177–183.
Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O'Hara, R. B.,

161

Simpson, G. L., Solymos, P., Stevens, M. H. H., and Wagner, H. (2015). vegan:
Community Ecology Package. R package version 2.2-1.
Palmer, C., Bik, E. M., DiGiulio, D. B., Relman, D. A., & Brown, P.O. (2007).
Development of the human infant intestinal microbiota. PLoS Biology, 5: e177.
Panda, S., El Khader, I., Casellas, F., López Vivancos, J., García Cors, M., Santiago, A.,
Cuenca, S., Guarner, F., & Manichanh, C. (2014). Short-term effect of antibiotics on
human gut microbiota. PLoS ONE, 9(4): e95476
Ringø, E., Zhou, Z., Vecino, J. L. G., Wadsworth, S., Romero, J., Krogdahl, J., Olsen, R.
E., Dimitroglou, A., Foey, A., Davies, S., Owen, M., Lauzon, H. L., Martinsen, L.
L., Bossier, P., Sperstad, S., and Merrifield, D. L. (2016). Effect of dietary
components on the gut microbiota of aquatic animals. A never-ending story?
Aquaculture Nutrition, 22: 219-282.
Romero, J., Feijoó, C. G., & Navarrete, P. (2012). Antibiotics in Aquaculture – Use ,
Abuse and Alternatives. In Carvalho, E., Gianmarco, S. D., Silva, R. J., editors
Health and Environment in Aquaculture, 160–198, InTech.
Romero, J., Ringø, E., & Merrifield, D. L. (2014). The Gut Microbiota of Fish. In
Merrifield, D. & Ringo, E., editors Aquaculture Nutrition: Gut Health, Probiotics
and Prebiotics, 75-100, John Wiley & Sons, Ltd., Chichester, UK.
Samira, S., & Guichard, B. (2006). Bacterial Antibiotic Resistance in Aquaculture. In
Didier, M. & Ramesh, R., editors Aquaculture Microbiology and Biotechnology,
133-157, Enfield: Science Publishers.
Schokker, D., Jansman, A. J. M., Veninga, G., de Bruin, N., Vastenhouw, S. A., de Bree,
F. M., Bossers, A., Rebel, J. M. J., & Smits, M. A. (2017). Perturbation of
microbiota in one-day old broiler chickens with antibiotic for 24 hours negatively
affects intestinal immune development. BMC Genomics, 18(241): 1-14.
Schokker, D., Zhang, J., Vastenhouw, S. A., Heilig, H. G. H. J., Smidt, H., Rebel, J. M.
J., & Smits, M. A. (2015). Long-Lasting Effects of Early-Life Antibiotic Treatment
and Routine Animal Handling on Gut Microbiota Composition and Immune System
in Pigs. PLOS ONE, 10(2): e0116523.
Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W. S., &
Huttenhower, C. (2011). Metagenomic biomarker discovery and explanation.
Genome Biology, 12(6): R60
Slama, T. G. (2008). Gram-negative antibiotic resistance: there is a price to pay. Critical
Care, 12(Suppl 4): S4.
Vadstein, O., Bergh, Ø., Gatesoupe, F.-J., Galindo-Villegas, J., Mulero, V., Picchietti, S.,

162

Scapigliati, G., Makridis, P., Olsen, Y., Dierckens, K., Defoirdt, T., Boon, N., De
Schryver, P., Bossier, P. (2012). Microbiology and immunology of fish larvae.
Reviews in Aquaculture, 5, S1–S25.
Vasoo, S., Barreto, J. N., & Tosh, P. K. (2015). Emerging Issues in Gram-Negative
Bacterial Resistance: An Update for the Practicing Clinician. Mayo Clinic
Proceedings, 90(3): 395–403.
Vellend, M. (2010). Conceptual synthesis in community ecology. The Quarterly Review
of Biology, 85(2):183–206.
Wickham, H. (2009). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag,
New York.
Winter, B. (2013). Linear models and linear mixed effects models in R with linguistic
applications. pp:1-22. Retrieved from [http://arxiv.org/pdf/1308.5499.pdf]
Zhao, L., Wang, G., Siegel, P., He, C., Wang, H., Zhao, W., Zhai, Z., Tian, F., Zhao, J.,
Zhang, H., Sun, Z., Chen, W., Zhang, Y., & Meng, H. (2013). Quantitative Genetic
Background of the Host Influences Gut Microbiomes in Chickens. Scientific
Reports, 3, 1163: 1-6

163

CHAPTER 4: BIOTIC AND ABIOTIC FACTORS INFLUENCING CHANNEL
CATFISH EGGS AND GUT MICROBIOME DYNAMICS DURING
EARLY LIFE STAGES
ABSTRACT
This study quantifies the effects of rearing environments associated with pond
aquaculture on microbial community composition in larval Channel Catfish (Ictalurus
punctatus) gastrointestinal (GI) tract across early developmental stages. Larvae were
exposed to hatchery water and nursery pond water environments and were sampled at
five developmental stages (Egg, Swim-up, Stocked, 24 hr post-stocking in nursery ponds,
and 21 d post-nursery pond stocking). Massive-parallel sequencing of a segment of the
16S rRNA gene was using an Illumina MiSeq platform to characterize the diversity and
taxonomic composition of gut and water microbial communities. PERMANOVA
analyses of differences in egg/gut community composition across stages indicated there
was a significant interaction between nursery pond environment, and family (p < 0.05)
during the last two stages. Several unclassified Proteobacteria and Firmicutes genera
along with other opportunistic pathogens like Vibrio, Aeromonas, Flavobacterium were
associated with differences in gut bacterial communities during larval development
before and after pond stocking. Our study demonstrate that rearing environment is an
important factor influencing the transfer of microbes from water or food into the gut, and
provides insight concerning the niche and ecological adaptability of microbes inside the
catfish gut. Findings suggest managers may have difficulty adopting probiotic treatments
to enhance gut microbiota in channel catfish aquaculture without appropriate pond
microbial management during early life stages.

164

INTRODUCTION
In the United States, channel catfish (Ictalurus punctatus; Rafinesque, 1818) and channel
x blue (Ictalurus furcatus; Valenciennes, 1840) hybrid catfish aquaculture are the largest
food fish commodities, listed at 162 million kg production in 2013 with an estimated
value of $354 million dollars (USDA NASS, 2013). In the past decades, catfish farming
has intensified to meet consumer demands. As a result, aquaculturists desire more
efficient land use and increased production in inland or pond aquaculture settings. These
improvements are essential to achieve successful and profitable catfish production (Avery
& Steeby, 2004; Hargreaves & Tucker, 2004). Part of these strategies involves research
associated with (but not limited to) feed efficiencies (Li, Manning, & Robinson, 2004; Li
et al., 2008), increased growth (Silverstein et al., 2000), water quality (Tucker &
Hargreaves, 2004) and enhanced disease resistance (Declercq et al., 2013; Park, Aoki, &
Jung, 2012).
Features of intensive fish cultivation such as grading, handling, transportation, as
well as poor water quality, may create conditions that are stressful to fish but favorable
for pathogenic microbes to thrive, thereby increasing the risk of disease (Derome et al.
2016, De Schryver & Vadstein, 2014; Skjermo & Vadstein, 1999). Catfish aquaculture
managers have emphasized creation and maintenance of proper living environments for
fish, including the design of facilities and operational production methods to minimize
the impacts of stress. Proper living environments include high-quality water within
tolerance limits of parameters such as temperature, dissolve oxygen; as well as stocking
density and feeding rates that is conducive to maximum performance (i.e,, best survival
growth and feed conversion) (Tucker, Avery, & Heikes, 2004).

165

However, even in well-designed and properly managed facilities, disease
outbreaks still occur, with approximately 45% of inventory losses attributed to infectious
disease (Hawke & Khoo, 2004). The complex interactions between fish and microbes
(Tucker & Hargreaves, 2004) during disease outbreaks are largely attributed to dysbiosis;
a phenomena associated with perturbations and imbalance in commensal microbial
communities present on hosts (Llewellyn et al., 2014; Karlsson et al. 2013). Disturbance
of natural microbial communities could happen due to acquisition of microbes from
shared, poor quality-water environment, or due to any environmental stressor that could
trigger negative physiological responses in fish.
In catfish aquaculture, the pond environment is usually characterized by a high
level of organic material. Biological activity in catfish ponds is stimulated by nutrient
loading derived from feeding (Tucker & Hargreaves, 2004). While this material is a
necessary component of catfish husbandry (by encouraging phytoplankton blooms which
produce oxygen), materials also support and can accelerate microbial growth. Fish feces
and unconsumed feed are important parameters that have to be carefully monitored
especially in recirculating water/ pond aquaculture systems. Organics materials can form
suspended particles that support growth of heterotrophic bacteria (Martins et al., 2013;
Skjermo & Vadstein, 1999). Physiochemical properties of water such as pH, salinity, and
dissolved organic carbon concentration can be adversely affected by many factors such as
weather, and as a result may trigger the growth pathogenic microorganisms (Martins et
al. 2013). The increasing number of harmful microbes could eventually overcome skin or
gill protective mechanisms, invading fish internal and vital organs, subsequently causing
disease outbreaks.

166

Next-generation sequencing technologies combined with advanced bioinformatics
capabilities have provided opportunities to expand studies of complex microbial
communities residing in fish hosts and in aquatic environments, resulting in great
potential for downstream applications in aquaculture management. Importantly, recent
studies have begun to focus on the importance of gut microbial communities
(microbiome/microbiota) in maintaining normal host functions (reviews in Ghanbari,
Kneifel, & Domig, 2015; Llewellyn et al., 2014; Nayak, 2010; Ringø et al., 2015).
Recognizing the critical roles played by gut microbiota has facilitated development of
probiotic applications in efforts to improve feed efficiency, expedite growth, and increase
disease resistance (Merrifield et al., 2010; Verschuere et al., 2000) in aquaculture
settings. However, an important first step in evaluating the potential of probiotics in
catfish aquaculture is to establish the role of genetics (family effects), husbandry
practices, and the environment on gut microbial community diversity and composition,
especially in pond based aquaculture as employed by the catfish industry in the
southeastern United States.
Previous studies have characterized the composition of microbiota associated with
several catfish species including Yellow Catfish (Pelteobagrus fulvidraco), Vietnamese
striped Catfish (Pangasianodon hypophthalmus Sauvage), and Amazonian freshwater
Catfish (Filhote ,Brachyplatystoma filamentosum and Dourada (Brachyplatystoma
rousseauxii) (Damasceno et al., 2016; Tong Thi et al., 2016; Wu et al., 2010, 2012),
focusing mostly on descriptive characterizations of microbial community composition.
Moreover, these studies were based on culture-based methods, or used molecular-based
methods for a specific stage of development.

167

Catfish producers are continually exploring production practices that will increase
feed conversion efficiency and growth. While the administration of pre- and probiotics to
beneficially manipulate the fish microbiomes is thought to improve growth and feed
efficiency in several fish species, probiotic applications in catfish culture are limited.
Moreover, information regarding the gut microbiota of catfish species and the factors
influencing them is limited (Carnevali et al. 2014). To maximize the potential of beneficial use of
microbes in catfish aquaculture, the

dynamics driving catfish gut microbial community

development and ontogenetic successional changes needs to be established, especially
during early life stages.
To our knowledge, only two studies have used molecular approaches to
interrogate channel catfish microbiomes, providing foundational work to molecularly
defined-channel catfish microbiota (Bledsoe et al., 2016; Larsen, Mohammed, & Arias,
2014). Production of catfish relies heavily on pond aquaculture. Given the likelihood of
compositionally differentiated microbial communities in different catfish production
ponds, the stability of the gut microbiome in resident fish populations raised in different
ponds needs to be determined. One fundamental but unresolved question is do catfish
populations raised in different environments possess the same gut microbial
compositions, or do gut microbial communities differ as a function of pond residency?
This is especially important when considering the role of probiotics, as catfish gut
microbial communities at some life stages and in particular pond environments may be
more conducive to probiotic manipulation than others.
In this study, our general objective was to characterize the taxonomic composition
and diversity of the gastrointestinal (GI) tract microbial community (microbiota) in

168

channel catfish across sequential ontogenetic stages from fertilized and incubating eggs
until 21 days following stocking into nursery ponds. In addition, comparisons of
microbial community composition and diversity were made from water from the hatchery
as well as from rearing pond. The study was designed to quantify associations between
host genotype (family/brood) and the rearing water supply, focusing specifically on the
effects of transitions from the hatchery to the nursery ponds on the gut microbial
community composition and diversity. Analyses quantified whether changes in gut
microbiota were influenced by family and/or by nursery pond. Our findings have general
implications for the management of fish nutrition, disease control, and probiotic use in
catfish pond aquaculture. Data will further reveal environmental or deterministic changes
during early ontogenetic stages and dynamic relationships between hosts and
environmental epibiota to which they are exposed.

169

METHODOLOGY
Fish husbandry and spawn collection
Channel catfish samples (eggs and fish) in this study originated from mated pairs from
spawning events in two outdoor 0.1-acre brood ponds (Ponds 147 and 154) on the same
day on May 29th, 2015. Brood ponds had been stocked with 60, three year old channel
catfish (1:1 female to male ratio) and maintained according to standard industry practices
(Tucker and Robinson 1990). Two discrete spawns/families were collected from each
brood pond. Upon collection, individual egg masses from spawns were treated with 100
ppm Povidone-iodine solution prior to transfer into separate tanks in the hatchery.
Treated eggs were maintained in discrete hatching tanks, supplied with flow-through well
water (~27°C; ~3.5 L/min) and constant aeration.
Upon hatching (June 5th, 2015), yolk sac fry were retained in the same hatching
tank and began exogenous feeding approximately 4 days post-hatch. Larvae were fed
trout starter diet (Rangen, Buhl, Idaho) until transfer to the nursery ponds at 19 days posthatch (dph). For each spawn, fry were stocked in net-pens (100 fish/pen), and placed in
two 0.1-acre nursery ponds. Nursery ponds were prepared according to standard
fertilization and pond preparation protocols (Mischke 2012). The schematic design for
this experiment in shown in Figure 4.1. Every two weeks, water samples were collected
and analyzed for chlorophyll a, ammonia-N, nitrite-N, and pH. Chlorophyll a was
determined using chloroform-methanol extraction (Lloyd and Tucker 1988). Ammonia
(Nesslerization) and nitrite (diazotization) were determined following the methods
outlined by the Hach Company (2013). Zooplankton and phytoplankton were assessed as
described previously (Tucker et al. 2017). In line with industry practices, feed was

170

withheld while fish were in net-pens, to encourage foraging on naturally occurring
zooplankton communities. The experiment ended 21-days post-stocking (~40 dph).
Respective spawns are referenced herein according to the corresponding brood identifiers
(147A, 147B and 154A, 154B).
Genomic DNA was isolated from 8 fry from each spawn and 9 microsatellite loci
were genotyped according to established methods (Waldbieser and Bosworth, 2103).
Parentage analysis using Cervus 3.0 (Kalinowski et al., 2007) revealed all fry from each
tank were a full-sib family belonging to one sire and one dam, and no parents were
shared between spawns. All fish husbandry was conducted at the Thad Cochran National
Warmwater Aquaculture Center in Stoneville, MS.
Sample collection
Fish were collected during five consecutive ontogenetic developmental stages, denoted as
Egg (fertilized egg samples in brood cans), Swim-up (fry collected during swim-up stage
in the hatchery, aged 4dph), Stocking (fry collected at stage in the hatchery just prior to
pond-stocking, aged 19dph), NP 24hr/Pond 24hr (fry collected 24hr after stocking into
nursery ponds, aged 20dph), and NP 21d/Pond 21d (fry collected 21d after stocking into
nursery ponds, aged 40 dph). Fish were euthanized using an overdose of MS-222
(250mg/L) and samples were transferred into 50 ml centrifuge tubes containing 95%
ethanol.
For water samples, 40 ml of pond water were collected and concentrated via
centrifugation (20,000xg) to obtain the pellet, which was resuspended in 40 ml of 95%
ethanol. Water samples were first obtained during egg incubation the same day eggs were
collected (denoted as Incubation), followed by sampling after egg hatch at the sac-fry

171

stage (Hatch). The remaining water samples were collected the same time fish were
sampled (designated as Swim-up, Stocking, NP 24hr/Pond 24hr, NP 21d/Pond 21d). All
fish and water samples were stored at ambient temperatures until dissection and bacterial
DNA extraction was performed.
Fish dissection
The intestine of each larval catfish was recovered from individual fish using aseptic
techniques. Exterior surfaces were swabbed with 100% ethanol before dissections of the
whole digestive tract using sterile instruments. The dissection were performed with slight
modification as previously described by (Milligan-Myhre et al., 2011). The intact GI
tracts were cut out from the fish body cavities, and the excised gut was immediately
transferred into filtered-sterilized 80% ethanol solution for DNA isolation. Due to the
small size of the gut, the composite of at least four tissue samples formed each technical
replicate, for each spawn within each brood pond (and from particular pond – if
applicable) at each time point. In addition to fish collected for molecular analysis,
additional fish (n=10) from the NP 24hr/Pond 24hr and NP 21d/Pond 21d samples were
subjected to gut content analysis following previously established protocols (Mischke et
al. 2003).
DNA extraction and 16S rRNA amplicon sequencing
Gut microbiota from channel catfish larvae were surveyed using high throughput
sequencing of the 16S rRNA gene. Each tube containing gut samples was first
centrifuged for 15min at 4oC to pellet tissues and bacteria before DNA was extracted.
The MoBio PowerSoil® DNA Isolation Kit (Carlsbad, CA, USA) including a beadbeating step was used following protocols for low-biomass samples as suggested by the

172

manufacturer with slight modification. The integrity of each DNA sample was assessed
based on the amplification of an approximately 1.4k bp of 16S rRNA gene (27F and
1389R) followed by gel agarose electrophoresis and DNA concentration were quantified
using Microplate spectrophotometer (BioTek®, Winooski,VT, USA).
Ninety-eight DNA samples (including two control samples) that have been
validated to contain sufficient bacterial DNA (as shown by the presence of amplicon
bands in electrophoresis) were subsequently submitted for sequencing at the Michigan
State University Research Technology Support Facility, RTSF (East Lansing, MI, USA).
All sequencing procedures, including the construction of Illumina sequencing libraries,
emulsion PCR, and MiSeq paired-end sequencing v2 platforms of the V4 region
(~250bp; primer 515F and 806R) followed standard Illumina (San Diego, CA, USA)
protocols. Michigan State’s Genomics RTSF (https://rtsf.natsci.msu.edu/genomics/)
provided standard Illumina quality control, including base calling by Illumina Real Time
Analysis v1.18.61, demultiplexing, adaptor and barcode removal, and RTA conversion to
FastQ format by Illumina Bcl2Fastq v1.8.4.
Sequence processing
Details pertaining to the sequence data analyses pipeline and computing
workflows were described as follows. Briefly, paired-end sequence merging, quality
filtering, “denoising”, singleton-sequence removal, chimera checking, taxonomic
assignments and Operational Taxonomic Unit (OTU) selection was conducted using an
open-source workflow based on methods implemented by program mothur v.1.36.1
(Kozich, Westcott, Baxter, Highlander, & Schloss, 2013). A reference-based OTU
clustering and taxonomic assignment was performed using the SILVA-based bacterial

173

reference database file provided in mothur to cluster sequences defined with 97%
identity. To minimize effects of under sampling while maintaining a broad dataset, the
final OTU table was rarefied to a depth of 2848 sequences per sample. Rarefaction
analyses were performed to equalize sampling coverage for all samples based on the
selected sequence depth. Three DNA samples with low sequence depth were discarded in
downstream analyses.
Analyses of bacterial community profiles and ecological statistics analyses
4.1 Alpha diversity
All measures of community diversity and similarity including Inverse Simpson (1/D)
diversity indices and OTUs richness within samples (alpha [α] diversity) were calculated
from the sequence data within mothur. To test which predictor variables were important
to explain the variation in alpha diversity indices, we fit different regression models
tested how the response variables (Inverse Simpson index and OTUs richness) were
affected by family, brood pond, and stages depending on the rearing location (brood
ponds, hatchery, nursery ponds).
To test for significant differences in diversity indices of egg samples between
brood ponds, samples were analyzed using a one-way ANOVA. To Analyze
compositional differences in fish gut samples at both swim-up and stocking stages from
different brood ponds, nested ANOVA analyses were performed on both Inverse
Simpson diversity index and taxa richness (family nested within brood pond). We further
tested whether pond effects were associated with variation in fish gut profiles based on a
three-way ANOVA. This analysis tested for significant differences in microbial
community diversity indices of fish gut samples between spawns from different stocking

174

ponds across stages, using only samples from the last two stages (after 24 hr and 21 d
post nursery pond stocking). The p-values less than 0.05 indicate significant differences
between pairwise comparisons. All analyses were performed based on information in the
summary files provided by mothur using the programming and statistical software, R
(version 0.98.978) base package.
4.2 Temporal and Differential abundance of OTUs
Custom R code written and implemented using packages plyr, dplyr and reshape2 was
used to calculate the relative abundance and determine identity of dominant taxonomic
phyla and taxa in all communities across sampling times (fish gut and water-associated
microbial communities). The relative abundance of all taxa within community samples
was calculated, and abundance at or exceeding 1% were considered dominant taxa
present across all OTUs. OTU relative abundances are displayed to show temporal
variability in fish gut community structure. Remaining taxa are grouped as ‘Others’.
4.3 Beta diversity
We used several packages implemented in R to perform comparative (beta [β]) diversity
analyses and ecological statistics on the OTU dataset. Briefly, we used vegan functions to
generate Bray-Curtis (BC) resemblance matrices (Oksanen, 2015) followed by the
cmdscale function to perform Principle Coordinate analyses (PCoA) based on BC
distance (Bray & Curtis, 1957). The ggplot and ggplots2 package (Wickham, 2009) were
then used to create ordination plots to visually compare gut bacterial community
composition and aquatic community composition as a function of different spawns and
ponds environment across sampling periods based on the three largest eigenvalues.
Multivariate hypotheses testing were conducted using the adonis function. We performed

175

multivariate analyses on the difference between locations of the centroids of each group
based on Permutational Multivariate Analyses of Variance (PERMANOVA) on
resemblance matrices (Anderson, 2001; Anderson et al., 2013). Tests were employed
because of the non-parametric and skewed nature of microbial community data.
4.4 Influence of spawn, pond effects on intestinal microbiota
To further analyze associations between spawn (families), brood ponds and nursery pond
(water) effects with fish egg/ gut microbiota, PERMANOVA analyses were performed
separately on Bray-Curtis distance matrices of fish associated microbial communities
based on developmental stages and/or rearing location (hatchery vs nursery pond) (Fig.
4.1). At the egg stage, comparisons were made between eggs from different brood ponds
using family/spawn as replicate due to lack of biological replication. Under the null
hypothesis, we did not expect groupings in egg community composition to differ
significantly based on predictor variables (e.g., brood pond).
For fry that were sampled during the next two developmental stages (swim-up and
stocking), their gut community compositions were analyzed in PERMANOVA based on
a nested analysis (family nested within brood ponds). A nested analysis was conducted
because fish that were produced from different mated pairs (spawns or families) but from
the same brood ponds were individually maintained in the same tanks inside the hatchery.
However, these fish were not exposed to the waterfrom another tank that contained the
other two spawns produced in another brood pond. PERMANOVA analyses were
performed to test for significant in groupings (centroid location) of gut microbial
community compositions between fish from different brood ponds (spawns) for each of
two developmental stages (swim-up and stocking).

176

After a brief rearing period in the hatchery, fish from each family were placed in
each 0.1 acre nursery pond. We conducted PERMANOVA analyses assuming under the
null hypothesis that gut community composition at each time point (after 24 hr and 21 d
post stocking), would not differ significantly among families, between brood ponds, and
nursery ponds. Whenever analyses revealed a pervasive influence of sampling time on
community composition, we performed a PCoA on the fish community for separate
stages using predictor variables based on ANOVA analysis to determine significant
differences between fish egg/gut communities attributed to each predictor variable.
4.5 Contribution of predominant taxa to sample communities’ composition
Contributions of individual taxa to whole-community profiles were represented using a
heatmap, constructed using the heatmap function from the Heatplus library in R. The
relative abundance of each taxa present in each sample (fish gut and water) was
calculated, and taxa whose relative abundance was less than 1% of at least one sample
were removed. The color scheme indicates the relative abundance of each OTU. The
dendogram was added based on average linkage hierarchical (UPGMA) clustering using
Bray-Curtis dissimilarity matrices to reveal phylogenetic patterns across samples.

177

RESULTS
Pond water quality, zooplankton and phytoplankton assessment
Water quality was in the desired ranges for typical of channel catfish nursery ponds
(Mischke et al. 2017). Mean water quality variables (SEM) were: nitrite-N 0.04 (0.019)
mg/L, ammonia-N 0.75 (0.37) mg/L, pH 8.51 (0.04), and chlorophyll a 146 (31.2) µg/L.
Zooplankton populations were abundant in the ponds, with desirable zooplankton for
channel catfish fry averaging over 450/L (Mischke, 2012). Phytoplankton populations
were dominated by Chlorophyta, Chrysophyta, and Cyanophyta.
Channel catfish gut contents and parentage determination
Gut contents revealed the channel catfish fry consumed predominantly copepods,
ostracods and cladocerans, depending on availability. This finding is consistent with
previous work on channel catfish fry feeding habits (Mischke et al. 2003). Meanwhile,
parentage analysis based on microsatellite marker determined that all four spawns were
produced from a single pair and all by different dams and sires.
Sequencing and sample summary
After filtering using mothur, our 16S rRNA amplicon dataset produced 3,564,729 high
quality reads. In total, we observed 4759 OTUs (3066 when omitting singleton OTUs)
defined at 97% sequence identity. From a total of 98 samples submitted for sequencing,
rarefication at 2848 sequences per sample eliminated three samples below this
sequencing coverage. Rarefaction analyses in mothur revealed that sequencing efforts
were exhaustive at a sequence depth of 2848 sequences per sample where total percent
coverage consistently exceeded 98%. We were able to sample a large portion of the

178

OTUs and diversity present while still retaining a large number of samples for fish of
each age.
Microbial community taxonomic composition
To characterize what phyla were present in the catfish gut community and taxa relative
abundance, we quantified the number of sequences represented by each phylum in all fish
samples over all developmental stages. We found that gut bacterial communities from all
samples were comprised of 29 microbial phyla. However, the most abundant phyla
including more than 95% of total sequences, in order of abundance included
Proteobacteria (28%), unclassified phylum (24.9%), Firmicutes (13.9%), Bacteroidetes
(12.4%), Actinobacteria (7.1%), Acidobacteria (4.2%), Verrucomicrobia (3.6%),
Planctomycetes (1.6%), Gemmatimonadetes (1.1%), Cholroflexi (0.8%), and
Fusobacteria (0.7%).
Phylum-level bacterial contributions to the gut microbiomes are shown according
to spawn and corresponding sampling time (Fig. 4.2a, 4.2b). Fish were maintained inside
the hatchery until they reached the swim-up stage (Fig. 4.2a), and later were stocked into
two different nursery ponds (pond 7 and pond 8; Fig. 4.2b). Microbiome composition
changed dramatically before nursery pond stocking (in the hatchery) and after nursery
pond stocking. Proteobacteria and several other phyla dominated gut communities prior
to stocking. 24 hrs after being transferred into nursery ponds, Firmicutes dominated gut
communities. The presence of members of the Fusobacteria phylum became apparent at
21 days post-stocking. Bacterial community profiles originating from water samples
generally resembled fish gut communities at the same developmental stages, although the
relative abundance of Proteobacteria fluctuated across sampling times. Firmicutes was

179

the pre-dominant phyla in water samples collected during post-stocking periods (Fig.
4.3a, 4.3b).
At the genus, several taxa from Proteobacteria, Unclassified phyla,
Bacteroidetes, Actinobacteria, and Firmicutes were consistently prominent throughout
fish developmental stages (data not shown). Ralstonia, Vibrio, and unclassified taxa from
Unclassified phyla were most consistent in relative abundance, however the unclassified
Proteobacteria are documented to occur in the highest percentage across stages, ranging
from 11% to 36% of total bacterial abundance. Two genera from the phylum
Fusobacteria (Cetobacterium and Fusobacterium) were most prevalent during later
stages after being stocked in the nursery pond (24hr and 21d post-release stages).
Microbial community diversity
Regression models using all predictor variables (family, brood pond, stages, pond
whenever applicable) did not reveal significant difference in in Inverse Simpson index.
However, tests for OTUs richness showed significant different between spawns at the egg
stage (df = 1, F = 242, and p-value < 0.01) and between stages (i.e., after 24 hr and 21 d
post-stocking) once fish were transferred into nursery ponds (df = 1, F = 6.485, p-value <
0.05).
Variation in gut bacterial community profiles in association with water and stages
To visualize relationships between bacterial community composition in the gut
and environmental (water) sources across fish developmental stages, PCoA was
performed to analyze samples in reduced dimensional space using ordination plots.
Variation in community membership among all environmental microbiota and fish gut
microbiota was influenced by age-dependent changes in the taxonomic composition of

180

prevalent bacterial in the larval catfish gut. Differences in community composition were
particularly prominent for samples collected before and after nursery pond stocking. In
contrast, considerable overlap was apparent across time among samples collected during
the first three sampling periods (Figure 4.5).
Statistical analyses of beta diversity based on Bray-Curtis distance across all
sampling periods testing for significant effects of developmental stage, D, brood pond
(B), spawn/family (F), and nursery pond (P) were displayed separately in Table 4.1, 4.2,
and 4.3, respectively. Eggs were analyzed separately (Table 4.1). We found no influence
of brood pond on egg microbial community composition (pseudo-F = 1.258, R2 = 0.386,
p = 0.333). We also found no significant difference in gut microbial community
composition across brood pond and stages (fish at swim-up and stocking) when fish were
reared in hatchery (Table 4.2). However, PERMANOVA comparing the last two stages
(after 24 hr and 21 d post stocking) revealed significant interactions (Table 4.3) involved
developmental stages, family and pond effects (PERMANOVA test pseudo-F = 1.863, R2
= 0.038, p < 0.05). We conclude that at least one pairwise combination of post-stocking
stages, family/spawn and pond significantly influenced the composition of fish gut
microbial communities.
To disentangling the effect of spawn from the pervasive influence of sampling
time on community composition, we performed the PCoA on the fish community for the
last two stages using family and pond. Analyses were followed with an ANOVA analysis
to compare fish gut communities using spawn as the predictor variable. We included
family and pond effects due to the significant interaction previously described. For each
stage, we were testing the null hypotheses that gut microbiome variations are not

181

significant between spawns. No significant effects (either pond or family) were detected
based on the independent ANOVA tests for the 24 hr and 21 d post nursery pond stocking
periods.
We next identified which genera contributed to gut microbiome variation across
sampling times. We performed hierarchical clustering to visualize associations among
microbial communities when samples (eggs, fish GI tracts, and water) were grouped
together. After rarefication to remove genera whose relative abundance was less than 1%
in at least one sample, a heat map was constructed (Fig. 4.6) from 76 pre-dominant
genera. The dendogram based on the unweighted pair group with arithmetic averaging
(UPGMA) clustering method showed that the community-clustering pattern based on
Bray-Curtis distance was concordant with PCoA community ordinations (Fig. 4.5).
Looking at nodes of dendogram to the left of the heatmap, bacterial communities from
water and fish gut samples collected at the same time cluster on the same branch. In
addition, within each sampling time, fish gut communities were more similar to each
other than to water samples from the same time.
Several unclassified taxa from Proteobacteria (Family Enterobacteriaceae) and
Firmicutes (Family Peptostreptococcaceae), including Clostridium sensu stricto persisted
across all samples and times. Interestingly, about 10 taxa (Caulobacter, Vibrio,
Unclassified Actinobacteria, Unclassified Bacteroidetes, Pseudomonas, Sphingomonas,
Staphylococcus, Ralstonia, Arthrobacter, Strenotrophomonas) can be grouped as
representative genera for samples from early time periods (Fig. 4.6). We also detected
taxa that usually associated with opportunistic fish pathogen such as Flavobacterium,
Vibrio, and Aeromonas to be present in high abundance (more than 1%) in either fish or

182

water samples. In addition, Edwardsiella was also detected in one water sample during
egg incubation stage (data not shown).

183

DISCUSSION
Channel catfish (Ictalurus punctatus) is the top pond-aquaculture species in the United
States with an estimated commercial value of over $350 million annually in US alone
(USDA NASS, 2013). However, disease outbreaks caused by primary opportunistic
pathogen such as Flavobacterium, Vibrio, Aeromonas, Edwardsiella often become the
limiting factor in this industry. Surveillance and early detection of these pathogens is
necessary for disease controls to avoid losses in profits to catfish producers. While
farmers usually rely on antibacterial treatments, recent and safer alternatives to combat
the disease are currently being widely studied focusing on monitoring healthy gut
microbiota and the application of probiotics. To date, only two molecularly defined
microbiota studies (Larsen et al., 2014; Bledsoe et al. 2016) have been conducted for
channel catfish.
Our findings corroborated the prevalence of Proteobacteria as found in the study
by Bledsoe et al. (2016) on channel catfish. Bledsoe et al. (2016) indicated
Proteobacteria was the most prevalent phyla in fish at 3 dph, and Firmicutes along with
Proteobacteria were the prevalent phyla when fish reached 65dph. In addition to
exploring changes in microbiome following fish developmental trajectories in control
environment, our study also evaluate whether gut microbial communities in fish from
different family stocked into different nursery ponds (approach that commonly employed
in industrial practice) experience ontogenetic shift in similar direction or being more
influenced from consequence of their pond environment. Similarly, we found
Proteobacteria dominated during early life stages and Firmicutes appeared after 24 hr
fish were stocked into the ponds. Larsen et al. (2014) conducted studies using three

184

commercial warmwater fish species (channel catfish, Ictalurus punctatus, largemouth
bass, Micropterus salmoides, and bluegill Lepomis macrochirus) that were maintained in
a fishing pond for recreational purposes and reported Fusobacterium (Genus
Cetobacterium) was among the most abundant taxa. This is consistent with our results
showing the appearance of the same taxa when fish were stocked into the nursery pond
(data not shown).
Our study identified putative deterministic (fish developmental trajectories) and
stochastic (dispersal of bacteria from surrounding water) factors that influenced catfish
intestinal microbiomes. Our results indicate that selection occurred within the host where,
specific taxa proliferated and appear to have been favored. However, substantial
variability in microbiome composition was documented over time (Fig. 4.2). The gut
community also resembled water sample communities that were collected during the
same time fish was sampled. Community characterizations revealed temporal shifts in
community membership and the possible influence of microbes originated from rearing
water supplies. In pond-aquaculture, fish that were stocked into nursery ponds were
allowed to feed naturally without artificial feeding (Larsen et al. 2014). Fish stocked into
nursery ponds at the time of exogenous feeding on resident pond fauna (including
zooplankton and macroinvertebrates available in the pond) appeared to have rapidly
acquired microbes from the pond community.
We found that the number of observed taxa (taxa richness) in egg bacterial
communities differed significantly between spawns but only during egg stage, suggesting
microgeographic variation in pond communities at the time of fertilization and initial
colonization of egg surface in containers where eggs were oviposited and fertilized.

185

Furthermore, the heat map (Fig. 4.6) characterizing clustering of egg communities within
the same branch compared to water samples collected during incubation shows that egg
communities are similar to water communities at the same time. Previous studies
(Fujimoto et al. 2013) have documented distinct egg bacterial communities in lake
sturgeon (Acipenser fulvescens) that change across early, middle, and late incubation
stages regardless of rearing water environment. In contrast, other literature describe egg
surface bacterial communities are usually colonized by ambient bacteria in the water as
fish spawn axenic eggs (Hansen&Olafsen, 1999; Verschuere et al. 2000).
As fish grow, their gut communities become more similar to other communities
from respective sampling point (regardless of family, brood pond origins) indicated by
grouping in Figure 4.5(a). We also conclude that environmental (water) communities are
important innocula to developing larvae especially during later stages when exogenously
feeding in ponds but not as much during early stages (Fig. 4.6). Statistical analyses on the
number of taxa also indicate significant differences, yet differences were only detected at
the final two stages (24 hr and 21 d following stocking into rearing ponds). Taking pond
effects into consideration, PERMANOVA detected a significant interaction between
effects of developmental stage, spawns, and ponds (for post-stocking sampling periods,
24 hr and 21d). Further analyses fitting linear regression models (two-way ANOVA) to
distinguish which pond had greater influence on variation in community composition
revealed neither pond had significantly greater effects in colonization of gut
communities. Post-hoc tests could be more conservative, whereas the significance effect
of pond was unresolved although significant interaction effects were detected by
PERMANOVA.

186

Some predominant taxa detected in the heatmap (Fig. 4.5) included Vibrio,
Aeromonas, and Flavobacterium, which are recognized as opportunistic fish pathogens
that have been associated with disease outbreaks. Other fish pathogens, Edwardsiella
ictaluri and Edwardsiella tarda are the causative agents of enteric septicaemia disease
and Edwardsiellosis in fish species including catfish and tilapia (Hawke et al., 1998;
Griffin et al., 2014, Reichley et al. 2015, Meye & Bullock, 1973). In our data, we have
detected the presence of taxa classified as Edwardsiella in the water samples during the
egg incubation stage (data not shown). However, prevalence in the community was low
(less than 1%). Studies have indicated that pathogens can survive in water columns for a
long period of time (Austin & Austin, 2012). F. columnaris for example can persist for
long periods in waters with high organic matter content (Wakabayashi, 1991).
Pathogenic Edwardsiella sp. are probably the most significant limiting pathogen
associated with first-year channel catfish fingerlings (Wise et al. 2004). The source of
this bacterium in catfish populations is unclear. It is thought the bacteria is already
present in the environment during early stages in the production cycle and might possibly
be part of the resident microbial population. Combinations of asymptotic fish carriers and
any environmental trigger later in production (Wise et al. 2004) could lead to disease
outbreaks. However, our analyses only identified Edwardsiella in the brood pond
environment and it was not detected in the fish or environment during the first 21 days of
production. This would imply E. ictaluri is absent or is present at levels below our limit
of detection (sample rarefication to 2848 sequences). While this suggests E. ictaluri may
not be present during these early stages of production, more research into the point source
of E. ictaluri to the resident population is warranted.

187

Previous studies have documented the applications of some microbial species
from genera documented in this study as aquacultural probiotics (Verschuere et al. 2000;
Merrifield et al. 2010). Target disease agents are usually bacterial, and fish infected by
pathogen have been shown to be succesfully treated in aquaculture (Llewellyn et al.,
2014). For example, Aeromonas hydrophila has been successfully used in vivo to treat A.
salmonicida infection in Oncorhynchus mykiss (Irianto and Austin, 2002). In different
study involving Vietnamese striped catfish (Pangasianodon hypophthalmus Sauvage),
strains of Bacillus were used as probiotics in in-vitro challenges against E. ictaluri and A.
hydrophila in controlled aquaria environments. These strains are mostly cultured from
soil and from intestine of channel catfish (Ran et al., 2012). Authors reported strains
conferred significant benefit in reducing stripped catfish mortality when used as
probiotics.
We have documented influences of pond environments on catfish gut microbiota.
Further investigations and comparisons of these findings relative to other fish species are
needed and may help to explain fish resilience and abilities to adapt to different culture
systems. Catfish ponds represent unique microcosms, and we have found that in terms of
microbial community composition, no two catfish ponds are alike. While some physical
and chemical aspects are similar from pond to pond, individual ponds possess unique
microbial and eukaryotic micro- and macro-invertebrate communities. In our hatchery
system, fish husbandry involved rearing fish in different environments throughout the
production cycle. The fact that catfish were artificially raised in different ponds through
the production cycle further complicated analyses, as no two catfish ponds are alike in
terms of resident microbial communities (Arias et al. 2006).

188

Initially, catfish were raised as single year classes to harvest in a single pond. To
meet demands for year-round products, a multi-batch system was adopted to replace the
previously employed “all-in, all-out” strategies. As a result, grow-out ponds are now
maintained with multiple year classes of fish and as the larger fish are harvested, they are
replaced with smaller fish. As a result, fry are stocked into nursery ponds and raised to
fingerlings or stockers. As they grow, individuals are split into separate ponds to be
grown to stocker size or are understocked as large fingerlings in grow-out ponds. Catfish
can be raised in as many as 3 to four different ponds over the course of the production
cycle, exposing them to multiple taxonomically diverse microbial communities.
The catfish in this study were artificially reared in different ponds, which
complicated analyses, although this more closely resembles what fish experience over the
course of production. Several caveats are also acknowledged in this study. We lacked
sufficient replication resulting in large variation in gut microbiomes among samples.
Variances decreased our ability to statistically demonstrate differences among treatments
(e.g spawn, pond).
Our work suggests the environment influences the gut community composition in
individual ponds that may influence levels of inter-pond heterogeneity in resident fish
disease resistance, feed conversion, growth, and other important production traits. Data
suggests that if the environment could be manipulated in such a way to select for
favorable gut communities, managers could potentially improve production efficiency in
catfish production. This would be considerably more challenging if the fish biota was
less dynamic or developed independent of environmental community organization.
Our work also suggests manipulation of the gut community through feed

189

administered pre- and probiotics may prove to be challenging, particularly when young
fish are feeding exogenously If pre- or probiotics are going to have a positive effect,
intervention should be administered at the hatchery stage where the environment can be
more readily altered. However, its unclear if a gut community developed at the hatchery
stage will persist once fish are stocked into ponds. It is difficult to manipulate the
environment in 10-acre catfish ponds to promote specific “beneficial” communities if the
processes that drive community selection are not well understood.
For future studies, we recommend more biological replicates be taken to quantify
the influences of treatments, including network analyses to decipher the microbe-microbe
interaction that might give insight to the pathway involve in disease development in fish
host. In addition to that, we could also apply community ecology theory to move beyond
descriptive towards applied studies (Adair & Douglas, 2017). This result could have
implications in future pond-aquaculture management strategies, perhaps involving
probiotics and prebiotics applications to combat disease.

190

APPENDIX

191

APPENDIX
Table 4.1 PERMANOVA analysis indicating lack of significant differences between
centroid location of egg microbial communities from different brood ponds
(PERMANOVA test pseudo-F = 1.258, R2 = 0.386, p = 0.333; permutation=24*)
Df
Brood pond (B)
Residuals
TOTALS

1
2
3

Sum
Sq
0.276
0.439
0.716

Mean
Sq
0.276
0.220

F-model

R2

Pr (>F)

1.258

0.386
0.614
1.000

0.333

*Maximum number of permutations generated due to small sample size

Table 4.2 PERMANOVA analysis indicating no significant differences between centroid
location of gut microbial communities were observed between fish from swim-up and
stocking developmental stages (D). Family (F: 147 A,147 B, 154 A, and 154 B) were
nested within brood pond (B: 147 vs 154).
Df
Dev.stages (D)
Brood pond (B)
Spawn/Family (F)
DxB
Residuals
TOTALS

1
1
1
1
19
23

Sum
Sq
0.286
0.470
0.300
0.330
7.916
9.302

Mean
Sq
0.286
0.470
0.300
0.330
0.417

192

F-model

R2

Pr (>F)

0.686
1.129
0.720
0.792

0.031
0.051
0.032
0.035
0.851
1.000

0.877
0.279
0.755
0.740

Table 4.3 PERMANOVA analysis indicating summarizing partitioning of variability
among fish gut microbiota from samples collected during 24 hr and 21 d post-release
into two different nursery ponds significantly influenced by developmental stages (D),
family (F), and nursery pond (P). Results revealed a significant interaction of these
factors influencing the microbial communities present in the gut (PERMANOVA test
pseudo-F = 1.863, R2 = 0.038, p < 0.05; permutation =1000)
Df
Dev.stages (D)
Brood pond (B)
Family/Spawns (F)
Rearing pond (P)
DxB
DxF
DxP
BxF
BxP
FxP
DxBxF
DxBxP
DxFxP
BxFxP
DxBxFxP
Residuals
TOTALS

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
31
46

Sum
Sq
1.250
0.236
0.187
0.421
0.227
0.293
0.376
0.222
0.205
0.211
0.210
0.264
0.546
0.230
0.315
9.078
14.269

Mean
Sq
1.250
0.236
0.187
0.421
0.227
0.293
0.376
0.222
0.205
0.211
0.210
0.264
0.546
0.230
0.315
0.293

193

F-model

R2

Pr (>F)

4.269
0.806
0.640
1.437
0.776
0.100
1.282
0.758
0.701
0.719
0.717
0.900
1.863
0.784
1.075

0.088
0.017
0.013
0.029
0.016
0.021
0.026
0.016
0.014
0.015
0.015
0.018
0.038
0.016
1.075
0.636
1.000

<0.001***
0.737
0.932
0.075
0.788
0.437
0.157
0.798
0.861
0.853
0.866
0.580
0.016*
0.765
0.022

HATCHERY

147

147A

147B

154A

154

154B

NURSERY POND

147A

147B

147A

147B

154A

154B

154A

154B

Legend
Can

147A Family

Brood pond

Nursery pond

Figure 4.1 Schematic design of the larval catfish microbiome study. Catfish samples
(eggs and fish) originated from eggs produced in mated pairs of channel catfish fish
from spawning events in two different outdoor 0.1 acre brood ponds (Ponds 147 and
154). Eggs from different spawns were transferred into the hatchery and maintained in
discrete hatching tanks (from incubation, hatch, swim-up, and stocked fry stages),
supplied with flow-through ground water. Larvae were transfer to the nursery ponds at
19 days post-hatch (dph). For each spawn, fry were stocked in net-pens (100 fish/pen)
and placed in two different 0.1 acre nursery ponds. Nursery ponds were prepared
according to standard fertilization and pond preparation protocols

194

(a)

(b)

Figure 4.2. Phyla level composition of bacterial communities identified from
channel catfish fry GI tracts (a) prior to stocking into nursery ponds (NP) and (b)
post stocking into nursery ponds. Relative abundance (percentages) of dominant
bacterial phyla found in gut microbiota of channel catfish fry from two spawns (brood
ponds 147 and 154) across five different developmental stages. Gut communities
collected after 24 hr (NP 24 hr) and 21 days (NP 21d) fish were stocked into nursery
pond were displayed for each of two ponds (p7 - pond 7 and p8 - pond 8). The most
prevalent six phyla based on relative abundance are shown in the bar chart
(Acidobacteria, Actinobacteria, Bacteroidetes, Firmicutes, Fusobacteria,

195

Figure 4.2 (cont’d)
Proteobacteria, Unclassified, Verrucomicrobia) and the remainder of taxa were
characterized as Others.
Caption (i)
Egg (fish egg samples);
Swim-up fry (fry collected during swim-up stage);
Stocked fry (fry collected during stage prior to pond-stocking)
NP 24hr/ Pond 24hr (fry collected at stage after 24hr being stocked into nursery pond)
NP 21d/Pond 21d (fry collected at stage after 24hr being stocked into nursery pond)

196

(a)

(b)

Figure 4.3 Phyla level compositions of bacterial communities identified from water
samples (a) prior to fish stocking into nursery ponds (NP) and (b) after fish
stocking into nursery ponds. Relative abundance (percentage) of dominant bacterial
phyla found in water samples associated with catfish fry husbandry across different
developmental stages. Gut communities collected after 24 hr (NP 24 hr) and 21 days
(NP 21 d) fish were stocked into nursery pond are displayed for each of two ponds (p7 pond 7 and p8 - pond 8). Samples from NP 21 d p7 was excluded from analyses due to
inadequate number of sequences. Only the top six phyla were shown in the bar chart
(Actinobacteria, Bacteroidetes, Chloroflexi, Firmicutes, Fusobacteria,
Proteobacteria,Unclassified) and the remainder of taxa were characterized as Others.

197

Figure 4.3 (cont’d)
Caption (ii)
Egg Incubation (water samples collected during egg incubation);
Egg Hatch (water samples collected once egg hatched and went through sac-fry stage);
Swim-up fry (water samples collected during fish swim-up stage);
Stocked fry (water samples collected during fish stage prior to pond-stocking)
NP 24hr/Pond 24hr (water samples collected at 24hr after fish being stocked into nursery pond)
NP 21d/Pond 21d (fry collected at stage at 21day after fish being stocked into nursery pond)

198

(a)

(b)

Figure 4.4 Estimates of Alpha diversity (Inverse Simpson index, (a); OTU richness
(Number of observed taxa), (b)) for channel catfish gut microbial communities
from all treatments across all developmental stages. Statistics comparing gut
microbiotas across treatment at different ages were calculated using ANOVA (Simple
one-way ANOVA for egg communities; Nested ANOVA –family within brood pond for
gut community composition of fish raised in hatchery; ANOVA for randomized design
for gut community composition of fish transferred into nursery pond). Each column with
error bar indicate mean estimate and standard error of diversity index, colored by rearing
pond (Hatchery, Nursery Pond – NP 7, and NP 8).

199

Figure 4.4 (cont’d)
(a) Alpha diversity in gut microbiota at each time point, as measured by Inverse
Simpson index.
(b) OTU richness based on number of taxa observed in gut microbiota from all treatment
across all time points.

200

(a)

(b)
Communities

Communities

egg

MDS with BC distance

egg

MDS with BC distance

0.2

water_hatch
swim−up

0.0

stocking
pond24hr

−0.2

pond21d
−0.4

water_incubation

0.4

PCoA3, Var explained=5.96%

PCoA2, Var explained=8.14%

water_incubation

water_hatch
0.2

swim−up
stocking

0.0

pond24hr
pond21d

−0.2

Type of communities

Type of communities
−0.4

Fish
−0.4

−0.2

0.0

0.2

PCoA1, Var explained=17.29%

−0.2

0.0

0.2

PCoA1, Var explained=17.29%

0.4

Water

0.4

Fish
Water

Figure 4.5 Visual representation of differences in gut microbiota of channel catfish fry (Bray-Curtis distances). Bray-Curtis
dissimilarities distance between communities originating from fish GI tracts and water samples for all developmental stages

201

(c)
Communities

egg

MDS with BC distance

water_incubation

PCoA3, Var explained=5.96%

0.4

water_hatch
0.2

swim−up
stocking

0.0

pond24hr
pond21d

−0.2

Type of communities
−0.4

−0.2

0.0

0.2

PCoA2, Var explained=8.14%

Fish
Water

Figure 4.5 (cont’d)
examined from both spawns are visualized using Principal coordinates analyses (PCoA) plots. Points represent a bacterial
community from each sample (a) Two dimensional representation of PCoA using first and second axis; (b) first and third axis; (c)
second and third axis.

202

Relative abundance

Post-stocking
Post-stocking

PriorPrior
to pond
to pond
stocking
stocking

Relative abundance

Figure 4.6 Genus-level contribution to gut microbiome community variation. Heat maps show the relative abundance value of

203

Figure 4.6 (cont’d)
each bacterial genus present in gut microbiota for each sample for all developmental stages. These values are depicted by color
intensity with the legend indicated on the top of the figure. A genera that could not be identified during the sequence alignment step
was categorized as unclassified, and the respective phylum information is provided. On the left of the heat map, each sample type was
distinguished into either fish-origin (denoted by filled circle) or water-origin (denoted by filled triangle).Different colors represent
different stages (or sampling point). The keys to each symbol and color is similar to Figure 4.5.

204

LITERATURE CITED

205

LITERATURE CITED
Adair, K. L., & Douglas, A. E. (2017). Making a microbiome: the many determinants of
host-associated microbial community composition. Current Opinion in
Microbiology, 35: 23–29
Anderson, M. J. (2001). A new method for non-parametric multivariate analysis of
variance, Australian Ecology, 26: 32-46
Anderson, M. J. (2006). Distance-based tests for homogeneity of multivariate
dispersions. Biometrics, 62: 245-253
Anderson, M. J. and Walsh, D. C. I. (2013). What null hypothesis are you testing?
PERMANOVA, ANOSIM, and the Mantel test in the face of heterogenous
dispersions. Ecology monographs, 83: 557-574
Arias, C. R., Abernathy, J. W., & Liu, Z. (2006). Combined use of 16S ribosomal DNA
and automated ribosomal intergenic spacer analysis to study the bacterial
community in catfish ponds. Letters in Applied Microbiology, 43: 287-292
Austin, B., & Austin, D. A. (2012). Bacterial Fish Pathogens Disease of Farmed and
Wild Fish (Fifth edition). Springer New York.
Avery, J. L., & Steeby, J. A. (2004). Hatchery Management. In Tucker, C. S. and
Hargreaves, J. A. (Ed.), Biology and Culture of the Channel Catfish, pp 145-165,
Elsevier B. V, Amsterdam, The Netherland
Bledsoe, J. W., Peterson, B. C., Swanson, K. S., & Small, B. C. (2016). Ontogenetic
Characterization of the Intestinal Microbiota of Channel Catfish through 16S rRNA
Gene Sequencing Reveals Insights on Temporal Shifts and the Influence of
Environmental Microbes. Plos One, 11(11): e0166379.
Carnevali, O., Sun, Y.Z., Merrifield, D. L., Zhou, Z., and Picchietti, S. (2014). Probiotic
Applications in Temperate and Warmwater Fish Species. In: Aquaculture Nutrition:
Gut Health, Probiotics and Prebiotics. D. Merrifield and E. Ringo, eds. Wiley
Blackwell
Damasceno, E. I. T., Pantoja, L. N. G., Figueiredo, H. M., Silva, L. H. M. & Rodrigues,
A. M. C. (2016). Microbiota of freshwater catfish species, Filhote
(Brachyplatystoma filamentosum) and Dourada (Brachyplatystoma rousseauxii)
from the Amazon Region (Belém-Pará-Brazil). African Journal of Microbiology
Research, 10(13): 445-455
De Schryver, P., & Vadstein, O. (2014). Ecological theory as a foundation to control
pathogenic invasion in aquaculture. The ISME Journal, 1–9.

206

Declercq, A. M., Haesebrouck, F., Van Den Broeck, W., Bossier, P., & Decostere, A.
(2013). Columnaris disease in fish: A review with emphasis on bacterium-host
interactions. Veterinary Research, 44(1): 1–17.
Derome, N., Boutin, S., Llewellyn, M., & Gauthier, J. (2016). Opportunistic pathogens of
fish. In Hurst C. J. (Ed). The Rasputin Effect: When Commensals and Symbionts
Become Parasitic. pp 649, Springer, New York.
Fujimoto, M., Crossman, J. A, Scribner, K. T., & Marsh, T. L. (2013). Microbial
community assembly and succession on lake sturgeon egg surfaces as a function of
simulated spawning stream flow rate. Microbial Ecology, 66(3): 500–511.
Ghanbari, M., Kneifel, W., & Domig, K. J. (2015). A new view of the fish gut
microbiome: Advances from next-generation sequencing. Aquaculture, 448: 464–
475
Griffin, M. J., Ware, C., Quiniou, S. M., Steadman, J. M., Gaunt, P. S., Khoo, L. H., and
Soto, E. (2014). Edwardsiella piscicida identified in the southeastern USA by gyrB
sequence, species-specific and repetitive sequence-mediated PCR. Diseases of
aquatic organisms 108(1): 23-35.
Hansen, G. H., & Olafsen, J. A. (1999). Bacterial Interactions in Early Life Stages of
Marine Cold Water Fish. Microbial Ecology, 38(1): 1–26.
Hargreaves, J. A., & Tucker, C. S. (2004). Industry Development. In Tucker, C. S. and
Hargreaves, J. A. (Ed.), Biology and Culture of the Channel Catfish, pp 1-14,
Elsevier B. V, Amsterdam, The Netherland
Hawke, J. P., Durborow, R. M., Thune, R. L., & Camus, A. C. (1998). ESC - Enteric
Septicemia of Catfish, Southern Regional Aquaculture Publication, (477), Retrieved
from http://www.aces.edu/dept/fisheries/aquaculture/pdf/477fs.pdf
Hawke, J. P. and Khoo, L. H., 2004. Infectious diseases. In Tucker, C. S. and
Hargreaves, J. A. (Ed.), Biology and Culture of the Channel Catfish, pp 347-443,
Elsevier B. V, Amsterdam, The Netherland
Hach Company. (2013). DR/3900 spectrophotometer procedure manual. Hach
Company, Loveland, Colorado.
Irianto, A. and Austin, B. (2002), Probiotics in aquaculture. Journal of Fish Diseases, 25:
633–642.
Kalinowski, S. T., Taper, M. L. & Marshall, T. C. (2007). Revising how the computer
program CERVUS accommodates genotyping error increases success in paternity

207

assignment. Molecular Ecology, 16: 1099-1106
Karlsson, F. H., Tremaroli, V., Nookaew, I., Bergstrom, G., Behre, C. J., Fagerberg, B.,
et al. (2013). Gut metagenome in European women with normal, impaired and
diabetic glucose control. Nature, 498, 99–103.
Kozich, J. J., Westcott, S. L., Baxter, N. T., Highlander, S. K., & Schloss, P. D. (2013).
Development of a dual-index sequencing strategy and curation pipeline for
analyzing amplicon sequence data on the miseq illumina sequencing platform.
Applied and Environmental Microbiology, 79(17): 5112–5120.
Larsen, A. M., Mohammed, H. H., & Arias, C. R. (2014). Characterization of the gut
microbiota of three commercially valuable warmwater fish species. Journal of
Applied Microbiology, 116(6): 1396–1404.
Li, M. H., Manning, B. B., & Robinson, E. H. (2004). Effect of daily feed intake on feed
efficiency of juvenile channel catfish. North American Journal of Aquaculture,
66(2): 100–104.
Li, M. H., Robinson, E. H., Peterson, B. C., & Bates, T. D. (2008). Growth and Feed
Efficiency of Juvenile Channel Catfish Reared at Different Water Temperatures and
Fed Diets Containing Various Levels of Fish Meal. North American Journal of
Aquaculture, 70(3): 347–352.
Llewellyn, M. S., Boutin, S., Hoseinifar, S. H., & Derome, N. (2014). Teleost
microbiomes: The state of the art in their characterization, manipulation and
importance in aquaculture and fisheries. Frontiers in Microbiology, 5 (207):1-17.
Lloyd, S. W., & Tucker, C. S. (1988). Comparison of three solvent systems for
extraction of chlorophyll a from fish pond phytoplankton communities. Journal of
the World Aquaculture Society, 19(2): 36-40.
Martins, P., Cleary, D. F. R., Pires, A. C. C., Rodrigues, A. M., Quintino, V., Calado, R.,
& Gomes, N. C. M. (2013). Molecular analysis of bacterial communities and
detection of potential pathogens in a recirculating aquaculture system for
Scophthalmus maximus and Solea senegalensis. PLoS ONE, 8(11): e80847
Meyer, F. P., and G. L. Bullock. (1973). Edwardsiella tarda, a new pathogen of channel
catfish (Ictalurus punctatus). Applied microbiology, 25(1): 155-156
Merrifield, D. L., Dimitroglou, A., Foey, A., Davies, S. J., Baker, R. T. M., Bøgwald, J.,
Castex, M., and Ringø, E. (2010). The current status and future focus of probiotic
and prebiotic applications for salmonids. Aquaculture, 302(1-2): 1–18.
Milligan-Myhre, K., Charette, J. R., Phennicie, R. T., Stephens, W. Z., Rawls, J. F.,
Guillemin, K., & Kim, C. H. (2011). Study of host-microbe interactions in zebrafish.

208

In Methods in cell biology, 105: 87-116
Mischke, C. C. (2012). Channel catfish pond fertilization. In Mischke, C. C (Eds).
Aquaculture pond ferttilization: Impacts of nutrient input on production, pp 137146, Wiley-Blackwell, UK.
Mischke, C. C., Griffin, M. J., Li, M. H., Wise, D. J., & Greenway, T. E. (2017).
Postponed feeding does not substantially reduce production expense during pond
rearing of hybrid catfish fry. North American Journal of Aquaculture, 79(2):135139.
Mischke, C. C., Wise, D. J., & Lane, R. L. (2003). Zooplankton size and taxonomic
selectivity of channel catfish fry. North American Journal of Aquaculture
65(2):141-146.
Nayak, S. K. (2010). Role of gastrointestinal microbiota in fish. Aquaculture Research,
41(11): 1553–1573.
Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O'Hara, R. B.,
Simpson, G. L., Solymos, P., Stevens, M. H. H., and Wagner, H. (2015). vegan:
Community Ecology Package. R package version 2.2-1.
Park, S. B., Aoki, T., & Jung, T. S. (2012). Pathogenesis of and strategies for preventing
Edwardsiella tarda infection in fish. Veterinary Research, 43(1): 1.
Ramette, A. (2007). Multivariate analyses in microbial ecology. FEMS Microbiology
Ecology, 62(2):142–160.
Ran, C., Carrias, A., Williams, M. A., Capps, N., Dan, B. C. T., Newton, J. C., … Liles,
M. R. (2012). Identification of Bacillus Strains for Biological Control of Catfish
Pathogens. PLoS ONE, 7(9): e45793
Reichley, S. R., Ware, C., Greenway, T. E., Wise, D. J., and Griffin, M. J. (2015). Realtime polymerase chain reaction assays for the detection and quantification of
Edwardsiella tarda, Edwardsiella piscicida, and Edwardsiella piscicida–like species
in catfish tissues and pond water. Journal of Veterinary Diagnostic Investigation,
27(2) 130-139.
Ringø, E., Zhou, Z., Vecino, J. L. G., Wadsworth, S., Romero, J., Krogdahl, J., Olsen, R.
E., Dimitroglou, A., Foey, A., Davies, S., Owen, M., Lauzon, H. L., Martinsen, L.
L., Bossier, P., Sperstad, S., and Merrifield, D. L. (2016). Effect of dietary
components on the gut microbiota of aquatic animals. A never-ending story?
Aquaculture Nutrition, 22: 219-282
Silverstein, J. T., Wolters, W. R., Shimizu, M., & Dickhoff, W. W. (2000). Bovine
growth hormone treatment of channel catfish: Strain and temperature effects on

209

growth, plasma IGF-I levels, feed intake and efficiency and body composition.
Aquaculture, 190(1–2):77–88.
Skjermo, J., & Vadstein, O. (1999). Techniques for microbial control in the intensive
rearing of marine larvae. Aquaculture, 177(1–4): 333–343.
Tong Thi, A. N., Samapundo, S., Devlieghere, F., & Heyndrickx, M. (2016). Microbiota
of frozen Vietnamese catfish (Pangasius hypophthalmus) marketed in Belgium.
International Journal of Food Contamination, 3(17): 1-9.
Tucker, C.C. and Robinson, E.H., 1990. Channel catfish farming handbook. Springer
Science & Business Media.
Tucker, C. S., Mischke, C. C., Brown, T. W., & Torrans, E. L. (2017). Water quality and
plankton communities in hybrid catfish (♀ channel catfish, Ictalurus punctatus x ♂
blue catfish, Ictalurus furcatus) ponds after partial fish harvest. Journal of the World
Aquaculture Society, 48(1): 46-56.
Tucker, C. S., Avery, J. L., & Heikes, D. (2004). Culture Methods. In Tucker, C. S. and
Hargreaves, J. A. (Ed.), Biology and Culture of the Channel Catfish, pp 166-195,
Elsevier B. V, Amsterdam, The Netherland
Tucker, C. S., & Hargreaves, J. A. (2004). Pond Water Quality. In Tucker, C. S. and
Hargreaves, J. A. (Ed.), Biology and Culture of the Channel Catfish, pp 215-278,
Elsevier B. V, Amsterdam, The Netehrland
USDA National Agricultural Statistic Services. (N.D. 2012). 2013 Census of
Aquaculture. Retrieved from https://www.agcensus.usda.gov/Publications/
2012/ Online_Resources/Aquaculture/Aqua.pdf
Verschuere, L., Rombaut, G., Sorgeloos, P., & Verstraete, W. (2000). Probiotic Bacteria
as Biological Control Agents in Aquaculture, 64(4): 655–671.
Waldbieser, G.C. and Bosworth, B.G. (2013). A standardized microsatellite marker panel
for parentage and kinship analyses in channel catfish, Ictalurus punctatus. Animal
genetics, 44(4): 476-479.
Wakabayashi, H. (1991). Effect of environmental conditions on the infectivity
of Flexibacter columnaris to fish. Journal of Fish Diseases, 14: 279–290.
Wickham, H. (2009). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag,
New York.
Wise, D. J., Camus, A. C., Schwedler, T. E., and Terhune, J.S. (2004). 15 Health
management. Developments in Aquaculture and Fisheries Science ,34: 444-503.

210

Wu, S., Gao, T., Zheng, Y., Wang, W., Cheng, Y., & Wang, G. (2010). Microbial
diversity of intestinal contents and mucus in yellow catfish (Pelteobagrus
fulvidraco). Aquaculture, 303(1–4): 1–7.
Wu, S., Wang, G., Angert, E. R., Wang, W., Li, W., & Zou, H. (2012). Composition,
diversity, and origin of the bacterial community in grass carp intestine. PloS One,
7(2): e30440

211