Factors that affect the composition and activity of salt marsh microbial communities: the effect
of nutrient loading and diel cycles
by Patrick J. Kearns B.S. in Biology, University of Massachusetts Boston
A dissertation submitted to
The Faculty of the College of Science of Northeastern University
in partial fulfillment of the requirements for the degree of Doctor of Philosophy
January 24, 2016
Dissertation directed by
Jennifer L. Bowen Associate Professor of Marine and Environmental Science
ii
Dedication
For mom, Jackie, dad, and A.J. for nurturing me and my curiosity. For John, Sarah, and Jen for getting me through it all.
For Ericka, my future.
iii
Acknowledgments I would sincerely like to thank all of the incredible people who have assisted me in not
only completing my dissertation, but also guiding me down the path that I am on today. First and
foremost, I would like to thank my advisor Jen Bowen for her guidance, compassion, friendship,
and her uncanny ability to always know what I wanted to hear or see to guide me along through
good and bad times. I’ve thoroughly enjoyed our journey that has been filled with good times,
good beer, amazing science, and conversations about science that have been among the most
exciting (and nerdy) of my life. I directly credit my development as a scientist to her mentorship
and I do not believe there are words to express how grateful and happy I am to have worked with
her for the past six years. In addition to Jen’s guidance, I would not be where I am today without
Rob Stevenson for giving me my first job in science, nurturing my curiosity, and always planting
new and novel ways to think about systems in my brain. I would also like to thank Mike Shiaris
for not only all his help over the years, but also setting me down the path of microbial ecology.
I’d also like to thank my Northeastern committee members Randall Hughes, David Kimbrio,
Katie Lotterhos, and Doug Woodhams for all their help and guidance.
While I have had many lab mates over the years, I would not have been able to complete
my work nor be the person I am today without the friendship and influence of John Angell and
Sarah Feinman. John and Sarah made graduate school bearable, fun, crazy at times, exciting,
food filled, and just an absolutely incredible experience. Sarah and John have profoundly shaped
my views of the world and their compassion and friendship have greatly changed me as a person.
In addition to John and Sarah I’d like to thank some particularly wonderful undergrads including
Dakota, Helen, Ian, Khang, Nhu, Maddie, Michael, Jen, and Sarah for their assistance in the lab
on parts of my dissertation and other projects. My ‘minons’ have shaped me into a better mentor
and have made me a more productive, focused person. Finally, I’d like to extend my gratitude to
Kenly Hiller and Ashley Bulseco-McKim for all their help, laughs, and making me think about
chemistry.
I have had the pleasure to collaborate with world-class scientists at WHOI, the MBL,
UMass Boston, Princeton, VIMS, and Villanova. In no particular order I’d like to thank Nat
West, Liam Revell, Marc Hensel, Jillian Dunic, Linda Deegan, Anne Giblin, David Johnson,
Jimmy Nelson, Laura Meyerson, Bess Ward, Andrew Babbin, Nick Peng, Tom Mozder, Rachel
Stanley, Amanda Spivak, Brandon LaBumbard, Mary O’Connor, Robin Elahi, Evan Howard,
iv
Jay Lennon, and Melanie Vile. These scientists have greatly expanded my knowledge of the
world around me and have made permanent marks on my scientific career. I’d like to give a
special thank you to Jarrett Byrnes for his amazing math mind, his willingness to always lend a
hand, and his sly ability to imprint the importance of statistics and programming permanently
into my brain. In addition to the scientists that I have interacted with and have shaped my
research, I would also like to extend my gratitude to those who made my work possible and
significantly easier through their actions. These people include Dave Behringer, Hillary
Morrison, David Dawson, Alexa McPherson, Rick Kesseli, Michie Yasuda, Thom Thera and
Illumina Inc., Yvonne Vaillancourt and her workers, Jeff Dusenberry, members of the TIDE
project, and members of the PIE-LTER.
Finally, I’d like to thank my family for their continued support and, at times, putting up
with freezers full of bugs. My mother for always making me think critically about any and
everything. For instilling science, steadiness, commitment, and reason into me from a very
young age and for molding me into the man I am today. Jaclyn for her friendship and support
through thick and thin. For instilling in me the patience of a saint and a strength to face down all
challenges, big or small. My father for always pushing me and encouraging me to be the best I
could be through hard work. Finally, I’d like to thank my fiancé Ericka for her love, compassion,
encouragement, and her smile. I look forward to our future filled with love, family, amazing
food, and sharing all of that with you mi amor. This dissertation embodies all the hard work and
confidence these people have instilled in me over the years.
v
Abstract of dissertation
Microorganisms, including bacteria and fungi, are the most abundant and species-rich
group of organisms on the planet. Their actions drive ecosystem processes and they confer key
ecosystem services that are of benefit to ecosystems worldwide. In addition to natural variation,
ecosystems are experiencing increased anthropogenic stressors which have profound effects on
both macro- and microscopic organisms alike. To better grasp how ecosystems will respond to
current and future changes we need to better understand how microbial communities and their
activity will be affected. In this dissertation I address the role of anthropogenic changes (excess
nitrogen) and natural changes (diel cycles) on microbial communities in salt marshes
ecosystems.
In chapter one, I examined how long-term fertilization (>10 years) alters bacterial
community composition and activity in salt marsh sediments. High-throughput sequencing of the
16S rRNA gene and 16S rRNA from sediments in two distinct salt marsh habitats (tall Spartina
alterniflora and S. patens) spanning nearly ten years of fertilization (2005, 2006, 2013, and
2014) revealed no effect of fertilization on overall bacterial community composition. The active
community, however, changed dramatically as a result of fertilization, resulting in a highly
homogenized active community that displayed lower active bacterial diversity and a higher
percentage of dormant taxa. My results suggest that dormancy is a prevalent strategy in marsh
sediments that can act to preserve microbial diversity in response to future environmental
change.
In chapter two, I analyzed the salt marsh fungal community response to excess nitrogen
across multiple habitats (mudflat, S. alterniflora, S. patens, and short S. alterniflora). Fungal
composition was altered by both nitrogen loading and habitat, with nutrient enrichment
significantly increasing both fungal diversity and abundance. Further, fertilization led to the
proliferation of taxa closely related to fungal denitrifiers, suggesting that nutrient enrichment
enhanced fungal denitrification. Organic matter in fertilized marshes showed greater signs of
decomposition, suggesting nitrate reduction contributed to decomposition of marsh peat and the
liberation of new carbon sources, which in turn promoted higher fungal diversity and biomass.
vi
My results highlight the important role of fungi to decomposition of marsh organic matter and
their response to long-term fertilization may have implications for carbon storage in marshes.
In my final chapter, I analyzed how both dynamic and static conditions in a salt marsh
pond affect composition and activity of bacteria over two diel cycles. Sediment communities,
which experience static conditions, displayed no net change in composition or abundance of both
the total and active community over the 48-hr experiment. Active communities in the pond
water, where abiotic conditions rapidly cycle on short time scales (<4hr), varied in concordance
with changes in their abiotic condition despite no net change in community composition. The
active pond water communities displayed a surprising dominance by members of the rare
biosphere (bacteria with <1% total abundance) suggesting rare taxa that possess the ability to
withstand large fluctuations in abiotic conditions are better adapted to those conditions and,
surprisingly, do not change the overall community composition. The results from my
dissertation highlight the influence of changes in abiotic conditions in structuring microbial
communities and their activity. Microbial communities mediate numerous ecosystem processes
and changes to the environment, both natural and abiotic, determine their composition.
Therefore, understanding how alterations to their composition, diversity, and abundance affect
their metabolic output will be key to understanding ecosystem response to future change.
vii
Table of Contents Dedication ii Acknowledgments iii Abstract of Dissertation v Table of contents vii List of figures ix List of tables x Introduction 1 Literature Cited 8 Chapter 1: Nutrient enrichment induces dormancy and decreases diversity of active bacteria in salt marsh sediment Abstract 13 Introduction 13 Methods 17 Results 25 Discussion 29 Acknowledgements 32 Literature Cited 32 Tables and Figures 37 Supplemental Materials 63 Chapter 2: Nutrient enrichment alters salt marsh fungal communities and increases decomposition of salt marsh sediments Abstract 48 Introduction 48
viii
Methods 50 Results 55 Discussion 57 Acknowledgements 61 Literature Cited 61 Tables and Figures 68 Supplemental Materials 76 Chapter 3: The effect of short-term, diel changes in environmental conditions of active microbial communities in a salt marsh pond Abstract 78 Introduction 78 Methods 81 Results 84 Discussion 88 Acknowledgements 91 Literature Cited 91 Tables and Figures 97 Supplemental Materials 104
ix
List of Figures Chapter 1 Figure 1. Effect of nitrogen enrichment on active and total bacterial community composition. 37 Figure 2. The effect of nitrogen enrichment on the percentage of dormant taxa and active bacterial diversity. 38 Figure 3. Order-level taxonomic composition of fertilized and reference sediments assessed by the 16S rRNA: 16S rRNA gene ratio. 39 Figure 4. Significantly different taxa between fertilized and reference sediments. 41 Figure 5. Effect of nutrient enrichment of net ecosystem metabolism. 42 SI Figure 1. The effect of fertilization of bacteria diversity in salt marsh sediments. 43 SI Figure 2. Titration of the 16S rRNA: 16S rRNA gene ratio for fertilized and reference sediments. 44 SI Figure 3. Weighted UniFrac similarity comparing different extraction and sample preservation methods. 45 Chapter 2 Figure 1. The effect of nutrient enrichment and habitat on fungal community composition. 69 Figure 2. The effect of nutrient enrichment and habitat on fungal abundance and diversity. 70 Figure 3. The effect of habitat and nutrient enrichment on taxonomic composition of fungal communities. 71 Figure 4. Significantly different fungal taxa between fertilized and reference marshes. 72 Figure 5. The effect of nutrient enrichment on the percentage of fungal denitrifiers. 73 Figure 6. Regressions of carbon against fungal community composition, diversity, and abundance. 74 Figure 7. The effect of nutrient enrichment and habitat on sediment carbon content, decomposition, and lability. 75 SI Figure 1. Schematic of a typical Plum Island marsh depicting the 4 habitats sampled. 76 SI Figure 2. The effect of habitat and nutrient enrichment on FTIR spectra. 77 Chapter 3 Figure 1. Environmental parameters in the pond sediment and water over the 48hr experiment. 97 Figure 2. The effect of diel cycles on total and active bacterial communities. 98 Figure 3. Changes in the percentage of active taxa in the sediment and water communities. 99 Figure 4. Taxonomic composition of active sediment and water communities over the 48hr experiment. 100 Figure 5. Heatmap of the 16S ratio for the most active water taxa of the 48hr experiment. 101 Figure 6. Plot of the relative abundance of the top 1000 OTUs in 16S rRNA relative to the 16S rRNA gene for sediment and water communities. 102 Figure 7. Plot of the 16S rRNA gene rank of the top 380 taxa versus the relative abundance in 16S rRNA for sediment and water communities. 103
x
SI Figure 1. qPCR results for sediment and water communities. 104 SI Figure 2. Active and total bacterial diversity for sediment and water communities. 105 SI Figure 3. Taxonomic composition of the total sediment community over the 48hr experiment. 106 SI Figure 4. Taxonomic composition of the total water community over the 48hr experiment. 107 SI Figure 5. Heatmap of the 16S ratio for the most active sediment taxa of the 48hr experiment. 108
List of Tables Chapter 1 SI Table 1. Results of Kruskal-Wallis test comparing the abundance of taxa between fertilized and reference sediments. 46 Chapter 2 Table 1. List of FTIR functional wavelengths. 68
1
Introduction
Microbes, including bacteria, fungi, other microeukaryotes, and viruses represent the
majority of biomass and species diversity on the planet (Hugenholtz et al. 1998). With millions
of species and well over 1025 individuals, microbes represent a dominant biological force
(Kallmeyer et al. 2012). Microbes play key roles in all aspects of ecosystems and play a
particularly important role in biogeochemical cycling of the elements essential for life, especially
carbon, sulfur, oxygen, and nitrogen (Falkowski et al. 2008). The actions of microbes in soils and
waters drive fluxes of key biological elements and fuel critical ecosystem processes, including
primary and secondary production (Falkowski et al. 2008). In addition to their role in nutrient
cycling and as ecosystem engineers, microbes also affect macroscopic life through intricate
symbioses (Mazmanian et al. 2008) and play key roles in host health (Cho and Blaser 2012).
While microbes are ubiquitous across the planet, microbial communities vary in both space and
time (Louzupone and Knight 2007) but these communities can also display surprising
consistency in their abundance and composition over decadal time periods (Kallmeyer et al.
2012, Kearns et al. 2016). The composition of microbes, in particular the set of genes a
community carries (metagenome), dictates the metabolic and ecosystem services microbes can
provide (Falkowski et al. 2008).
Studies examining microbial communities demonstrate a sensitivity of microbes to
perturbations in their environment (Allison and Martiny 2008) and their sensitivity to
environmental change is likely due to the narrow niche space most microbes occupy (Kirchman
2015). Studies examining the structure-function relationship of microbial communities has
revealed mixed results. Work by Cavigelli and Robertson (2000) and Reed and Martiny (2013)
demonstrate a tight relationship between the structure or change in structure of microbial
communities and the functional output of those communities. These studies are similar to those
done with macroscopic communities that report structure-function relationships (Thompson et al.
2012). While several studies demonstrate a link between microbial composition and function,
numerous other studies show no net effect of environmental change on function despite changes
in community composition (Bottomley et al. 2004, Balser & Firestone 2005, Langenheder et al.
2005). The disconnect between structure and function is typically ascribed to the presence of
functional redundancy in microbial communities. Functional redundancy allows microbial
communities to perform identical functions despite changes in community structure, likely due to
2
the overlapping metabolic capabilities of microbes occupying the same physical space (Shade et
al. 2012). While microbial communities can change in response to perturbations, communities
can display resistance to change despite changes in geochemical output, as demonstrated in salt
marsh sediments (Koop-Jakobsen and Giblin 2010, Bowen et al. 2011). When communities are
resistant to environmental change their composition remains unperturbed, but their function may
change due to changes in their active communities. The ability of microbial communities to
display resistance and resilience to change is commonly conferred by lower abundance taxa that
occupy similar niche space as dominant taxa allowing them to adapt to novel environmental
conditions (Lynch and Neufeld 2015).
In most habitats, microbial communities are typically dominated by only a few highly
abundant taxa (<10), despite their diversity (Sogin et al. 2006, Pedrós-Alió 2012). Thus,
microbial communities typically contain a large number of low abundance taxa, collectively
known as the rare biosphere, with abundances less than one percent in the total community
(Sogin et al. 2006; Lynch and Neufeld 2015). While macroscopic communities typically contain
organisms specifically adapted to local conditions, in the rare biosphere microbial communities
can contain taxa with metabolic and physiological requirements drastically different than would
be predicted based on local conditions. For example, Wissuwa et al. (2016) isolated thermophilic
bacteria from arctic tundra and sediments, where temperatures rarely achieve the optima for
thermophilic growth. Rare taxa are important contributors to activity of bacterial communities in
coastal waters (Campbell et al. 2011) and their enhanced activity can contribute to pulses of
ecosystem function under different abiotic conditions (Aanderund et al. 2015). Further, rare taxa
can become more abundant under conditions conducive to their growth and enhanced cell
division can significantly alter community composition (Shade et al. 2014). Overall most rare
taxa remain inactive or dormant and act as a seed bank of microbial diversity (Lennon and Jones
2011) that can respond to future changes in the environment. While dormant taxa such as
sporulated Firmicutes can remain dormant for hundreds of years (de Rezende et al. 2013, Müller
et al. 2014), the dormancy and activity of taxa is often cyclical or occasional as conditions
change within ecosystems (Salazar-Villegas et al. 2016). The bank of inactive taxa allows
microbial communities to respond to new conditions and to maintain ecosystem function,
allowing for the preservation of ecosystem services under stressful conditions.
3
Microbial dormancy is a bet-hedging strategy microbes employ to help them persist in
the environment under unfavorable environmental conditions or conditions not conducive to
their growth or reproduction (Lennon and Jones 2011). Dormancy is a strategy used by all three
domains of life, however, the importance of dormancy in eukaryotic lineages appears to be
significantly less than bacterial lineages (Jones and Lennon 2010). The proportion of dormant
taxa varies across ecosystems (20-50% of taxa) and their ubiquity can impede our ability to
detect changes in microbial communities and to relate community composition to associated
changes in ecosystem function. The controls over dormancy in ecosystems are poorly
understood, however dormancy can be a response to nutrient limitation in temperate lakes (Jones
and Lennon 2010) and changes in abiotic factors such as salinity can affect percentages of
dormant taxa (Aanderund et al. 2016). Further, pulses of ecosystem limiting factors, such as
water to desert soils, increases activity and perhaps alleviates dormancy (Frossard et al. 2015).
To assess levels of dormancy and document changes in active microbial communities, recent
work targets 16S rRNA in addition to the 16S rRNA gene (e.g. Campbell et al 2011). 16S rRNA
is used to assess dormancy because production of 16S rRNA tends to be positively correlated
with protein synthesis and thus activity of microbial communities (Kerkhof and Kemp 1999,
Deutscher et al 2006). Although there are some caveats in using this approach to directly predict
microbial growth (Blazewicz et al 2013), it is a reasonable proxy for protein synthesis potential,
indicating either microbial growth or other metabolic activity.
My dissertation examines the effect of changes in abiotic conditions on microbial
communities in a salt marsh complex in the Plum Island Ecosystems Long-Term Ecological
Research Site in Rowley, MA, USA. Salt marshes are critically important coastal ecotypes lying
at the interface of the land and the sea. Marshes are among the most productive ecosystems on
the planet (McLeod et al. 2011) and due to their high levels of primary productivity they provide
numerous ecological and economic services, including carbon sequestration, coastal protection
from storm surges, habitat for numerous species of animals, and removal of anthropogenic
nutrients (Millennium Ecosystem Assessment 2005). However, despite their importance,
marshes have experienced significant declines in recent decades likely due to anthropogenic
stressors including sea level rise, excess nutrients, alterations to hydrology, and invasive species
(Craft 2007, Craft et al. 2009, Deegan et al. 2012, DeLaune and White 2012). These ecosystem
stressors can alter the ability of a marsh to deliver key ecosystem services (e.g. remove N; Drake
4
et al. 2009), however, there are many unknowns about how marshes respond to global changes,
in particular, how their associated microbial communities respond to environmental change.
With densities of microbes of 109 per gram of sediment, marshes represent a substantial reservoir
of microbial diversity (Rublee and Dornseif 1978, Goodfriend 1988, Bowen et al. 2011), and
microbial activity affects rates of carbon sequestration and remineralization (Kirwan and Blum
2011), nutrient removal (Koop-Jakobsen and Giblin 2010), and the growth and function of plants
(Philippot et al. 2013; Hughes et al. 2014). Because of their instrumental role in salt marsh
ecosystem processes, understanding how stressors, both anthropogenic and natural, affect
microbial community structure and function is paramount to understanding how these
ecosystems will respond to future global change.
In chapter one I examine how long-term fertilization influences the composition and
activity of salt marsh bacterial communities. Numerous studies examining the influence of
excess nitrogen on microbial communities demonstrate a significant effect of excess nutrients
(Alison and Martiny 2008, Leff et al. 2015). Salt marsh microbial communities, however, are
resistant to excess nutrients (Bowen et al. 2009a, 2011) despite the observed alteration to
geochemical cycling (Koop-Jakobsen and Giblin 2010), activity (Bowen et al. 2009b), and
below-ground decomposition (Deegan et al. 2012). I hypothesized that the lack of response of
the total microbial community to excess N was due to high rates of dormancy present in marsh
sediments. Further, I hypothesized that long-term fertilization would alter the active microbial
community, leading to a homogenized group of active taxa. To test my hypotheses, I sequenced
the 16S rRNA gene and 16S rRNA to assay the total and active microbial communities
respectively. I sequenced nucleic acids from sediment samples collected monthly (May-October)
from 2005-2006 and 2013-2014 from two distinct marsh habitats (tall Spartina alterniflora and
S. patens) in fertilized (n=2) and reference (n=2) marshes. Consistent with previous work
(Bowen et al. 2011), my results suggest that fertilization does not alter the total microbial
community, despite over ten years of exposure to excess N. The active community, however,
was altered by fertilization, leading to an increase in active taxa that are better adapted to higher
N loads, including the overwhelming domination of Cyanobacteria and one specific sulfate
reducing bacteria (order Desulfobacterales). On average, marshes contain a large reservoir of
dormant taxa (~55%), which is often found in systems with rapidly changing abiotic conditions,
suggesting that dynamic environmental conditions promote dormancy as an ecological strategy.
5
Fertilized sediments, by contrast, displayed a large increase in dormant taxa (up to 90%), which
resulted in lower active microbial diversity. My results suggest that despite the resistance of the
total microbial community to change by excess N, the active community, and therefore the
metabolic output of marsh microbes, is dramatically altered due to the activity of a small number
of taxa. The dormancy of taxa in fertilized creeks allows for the maintenance of bacterial
diversity better able to respond to future global change.
In chapter 2, I followed up my analysis of how nutrient supply alters bacterial community
structure by examining the role of long-term fertilization in structuring fungal communities, their
diversity, and abundance across different marsh habitats. Terrestrial fungal communities can be
structured by carbon availability (Tedersoo et al. 2014) and excess nitrogen (Leff et al. 2015),
with excess nitrogen typically decreasing the diversity of fungi (Lin et al. 2012). Marine fungi,
particularly in coastal habitats, are poorly characterized both taxonomically and functionally
(Hawksworth, and Rossman 1997, Picard 2017), and we lack a fundamental understanding of the
factors that structure fungal communities. To better understand these factors, I examined fungal
communities and their response to excess nitrogen across four distinct marsh habitats (mudflat,
tall S. alterniflora, S. patens, short S. alterniflora; Fig. 1) using high-throughput sequencing and
quantitative PCR (qPCR) of the fungal internal transcribed spacer region (ITS).
Figure 1- A schematic of a typical marsh in the Plum Island Ecosystems marsh complex. Nitrogen loads (g N m-2 y-1) are indicated for reference (blue) and fertilized (red) marshes below
each habitat. Modified from Johnson et al. (2016). MHW=mean high water. MLLW=mean lower low water.
6
In addition to analyzing fungal abundance and community structure, I also used Fourier
Transform Infrared Spectroscopy (FTIR) to analyze organic matter quality. I hypothesized that
fungal communities would vary as a function of habitat, and that nutrient enrichment would have
no net effect on fungal communities as has been previously shown for marsh bacterial
communities (Bowen et al. 2009, 2011). Further, I hypothesized that excess nitrate would favor
the use of nitrate for anaerobic metabolism, resulting in enhanced decomposition of marsh
organic matter. My results indicate that fungal community composition, diversity, and abundance
varied by habitat in both fertilized and reference marshes. In contrast to results from terrestrial
soils (Lin et al. 2012), however, fertilization, significantly increased fungal diversity and
abundance, especially in low marsh habitats. Analysis of the taxonomic composition of fertilized
sediments revealed that many of the fungal taxa responding to excess N are closely related to
taxa that are denitrifying fungi (Maeda et al. 2015), suggesting that the increased supply of
nitrate likely stimulated fungal denitrification. FTIR analysis revealed that regions of the marsh
receiving the greatest exposure to nitrate (mudflat and tall S. alterniflora) displayed enhanced
decomposition and decreased lability of residual carbon, while carbon in the high marsh (S.
patens and short S. alterniflora) was not different between reference and nutrient enriched
marshes. My results highlight the important role of habitat and the associated abiotic conditions
in structuring fungal communities and their importance in decomposition of organic matter in
marsh sediments. Fertilization altered fungal communities through the proliferation of taxa that
are likely able to use nitrate as an anaerobic electron acceptor or that are more capable of
assimilating nitrogen into biomass. The enhanced use of nitrate in anaerobic metabolisms
resulted in enhanced decomposition in low marsh habitats (mudflat and tall S. alterniflora),
which likely resulted the release of additional carbon sources allowing for more fungal biomass
and higher fungal diversity. Further, the addition of nitrate did not alter carbon composition in
the high marsh habitats (S. patens and short S. alterniflora) suggesting primary producers replace
the decomposed carbon through their substantial root biomass.
In my final chapter I examine the role dynamic conditions play in the activity of bacterial
communities in a salt marsh pond. Salt marsh ponds are shallow ponds that form on the marsh
platform and are likely increasing as a result of anthropogenic sea level rise (Raposa et al. 2016).
Marsh ponds are typically isolated from tidal flooding except during the highest tides. During
their isolation, the water in marsh ponds experience dynamic changes in abiotic conditions on
7
short time scales (<4 hr) while the sediments remain stable. Lennon and Jones (2011)
demonstrated that percentages of dormant taxa are higher in systems that are both spatially and
temporally variable. Thus, I hypothesized that the dynamic conditions in the pond water would
promote dramatic changes in bacterial activity, while the stable conditions in the underlying
sediment would promote stable active bacterial communities. Further, I hypothesized that
dynamic conditions would promote greater inactivity overall, but would lead to enhanced
activity of taxa better adapted to conditions in the pond. I demonstrate that while both the total
and active communities remained stable in the sediments, active pond water communities
dramatically changed over short time scales with no net change in the total community. In the
water the most active taxa were rare taxa (<1% total abundance) demonstrating a surprising role
of the rare biosphere in the activity of the pond community. Finally, the percentage of inactive
taxa in the water (>55%) was higher than the sediment (~50%) suggesting that changing
conditions within the overlying water drive the majority of taxa into inactivity, including
numerically dominant taxa. The results from my third chapter underscore the rapid pace at which
microbial communities can respond to changes in their environment as well as the importance of
abiotic conditions in affecting the inactivity of microbial taxa. Further, the overwhelming
abundance of rare taxa in the active pond water communities and the high percentage of inactive
taxa suggests that rare taxa may be better adapted to dynamic conditions and the use of
dormancy may be higher in dynamic systems.
Microbial communities are a ubiquitous group of taxa that exert powerful controls over
ecosystem processes worldwide. Microbial communities are sensitive to changes in
environmental conditions (Alison and Martiny 2008) and changes in their composition and
activity can have profound effects on ecosystem function. While microbes are important
contributors to ecosystems, a substantial proportion of microbial taxa can be inactive and
therefore not contributing to ecosystem function. My dissertation examined the interplay
between environmental change, dormancy, community composition, and activity of microbial
taxa. Long-term nutrient enrichment differentially affected bacterial and fungal communities in
salt marsh sediments. Fungal community composition was changed as a result of increased
anthropogenic nitrogen while bacterial communities remained resistant, despite changes in
activity. The discrepancy between bacterial and fungal lineages may be due the differential use
of dormancy between bacteria and fungi (Jones and Lennon 2010). Conversely, fungi may be
8
more sensitive to environmental change than bacteria. Despite the domain-level differences in
response to N, the changes in activity of bacterial and fungal communities due to excess N has
led to a drastically altered marsh landscape. Fertilized marshes display an enhanced degree of
decomposition and accelerated carbon cycling relative to reference marshes. Further, my results
indicate a switch in the primary anaerobic electron acceptor from sulfate to nitrate in fertilized
creeks resulting in enhanced decomposition, which has been previously documented (Koop-
Jakobsen and Giblin 2010). The altered carbon cycle in fertilized marshes likely alter the ability
of the marsh to store carbon and the enhanced decomposition of marsh peat may be altering
marsh geomorphology (Deegan et al. 2012). The enhanced decomposition of marsh organic
matter is tied to the increase in activity and abundance of heterotrophic bacterial and fungal taxa
better adapted to higher nitrogen, which suggests a domain-independent ability for enhanced
growth under higher nitrogen loads.
My work highlights the important role of dormancy to salt marsh bacterial communities.
Salt marshes are highly dynamic ecosystems, experiencing changes in abiotic conditions on very
rapid time scales. The dynamic conditions appear to promote dormancy as an ecological strategy,
allowing communities to remain stable during perturbations and over decadal time scales.
Further, dormancy acts as a mechanism to preserve bacterial diversity under environmental
change. Salt marshes provide critical ecosystem services to coastal areas and their ability to do so
is directly tied to their microbial communities. My work links dormancy and changes in fungal
communities to alterations of marsh carbon cycling under nitrogen enrichment and the long-term
nitrogen supply to coastal systems may profoundly affect the ability of marshes to store carbon.
Despite the constant press of anthropogenic perturbations received by coastal systems, dormancy
of bacterial communities allows for the preservation of species and metabolic diversity, allowing
communities to continue to provide critical ecosystem services under current and future global
change.
Literature Cited
Aanderud ZT, Jones SE, Fierer N, Lennon JT (2015). Resuscitation of the rare biosphere
contributes to pulses of ecosystem activity. Front Microbiol 6: 24.
9
Allison SD, Martiny JB (2008). Resistance, resilience, and redundancy in microbial
communities. Proc Natl Acad Sci USA 105: 11512-11519.
Balser TC, Firestone MK (2005). Linking microbial community composition and soil processes
in a California annual grassland and mixed-conifer forest. Biogeochemistry 73: 395-415.
Blazewicz SJ, Barnard RL, Daly RA, Firestone MK (2013). Evaluating rRNA as an indicator of
microbial activity in environmental communities: limitations and uses. ISME J 7: 2061-2068.
Bottomley PJ, Taylor AE, Boyle SA, et al. (2004). Responses of nitrification and ammonia-
oxidizing bacteria to reciprocal transfers of soil between adjacent coniferous forest and meadow
vegetation in the Cascade Mountains of Oregon. Micro Ecol 48: 500-508.
Bowen JL, Crump BC, Deegan LA, Hobbie JE (2009). Salt marsh sediment bacteria: their
distribution and response to external nutrient inputs. ISME J 3: 924-934.
Bowen JL, Ward BB, Morrison HG, Hobbie JE, Valiela I, Deegan LA, Sogin ML (2011).
Microbial community composition in sediments resists perturbation by nutrient enrichment.
ISME J 5: 1540-1548.
Campbell BJ, Yu L, Heidelberg JF, Kirchman DL (2011). Activity of abundant and rare bacteria
in a coastal ocean. Proc Natl Acad Sci USA 108: 12776-12781.
Cavigelli MA, Robertson GP (2000). The functional significance of denitrifier community
composition in a terrestrial ecosystem. Ecology 81: 1402-1414.
Cho I, Blaser MJ (2012). The human microbiome: at the interface of health and disease. Nat Rev
Genet 13: 260-270.
Craft C (2007). Freshwater input structures soil properties, vertical accretion, and nutrient
accumulation of Georgia and US tidal marshes. Limnol Oceanogr 52: 1-11.
Craft C, Clough J, Ehman J, et al. (2009). Forecasting the effects of accelerated sea-level rise on
tidal marsh ecosystem services. Front Ecol Environ 7: 73-78.
DeLaune R, White J (2012). Will coastal wetlands continue to sequester carbon in response to an
increase in global sea level?: A case study of the rapidly subsiding Mississippi river deltaic plain.
Clim Chang 110: 297-314.
Deutscher J, Francke C, Postma PW (2006). How phosphotransferase system-related protein
phosphorylation regulates carbohydrate metabolism in bacteria. Microbiol Mol Biol Rev 70: 939-
1031.
10
Falkowski PG, Fenchel T, Delong EF (2008). The microbial engines that drive Earth's
biogeochemical cycles. Science 320: 1034-1039.
Frossard A, Ramond JB, Seely M, Cowan DA (2015). Water regime history drives responses of
soil Namib Desert microbial communities to wetting events. Sci Rep 5: 12263.
Goodfriend W (1998). Microbial community patterns of potential substrate utilization: a
comparison of salt marsh, sand dune, and seawater-irrigated agronomic systems. Soil Biol
Biochem 30: 1169-1176.
Hawksworth DL, Rossman AY (1997). Where are all the undescribed fungi? Phytopathology 87:
888e891.
Hugenholtz P, Goebel BM, Pace NR (1998). Impact of culture-independent studies on the
emerging phylogenetic view of bacterial diversity. J Bacteriol 180: 4765-4774.
Hughes AR, Moore AF, Piehler MF (2014). Independent and interactive effects of two
facilitators on their habitat-providing host plant, Spartina alterniflora. Oikos 123: 488-499.
Jones SE, Lennon JT (2010). Dormancy contributes to the maintenance of microbial diversity.
Proc Natl Acad Sci USA 107: 5881-5886.
Kallmeyer J, Pockalny R, Adhikari RR, et al. (2012). Global distribution of microbial abundance
and biomass in subseafloor sediment. Proc Natl Acad Sci USA 109: 16213-16216.
Kearns PJ, Angell JH, Howard EM, Deegan LA, Stanley RH, Bowen JL (2016). Nutrient
enrichment induces dormancy and decreases diversity of active bacteria in salt marsh sediments.
Nat Commun 7: 12881.
Kerkhof L, Kemp P (1999). Small ribosomal RNA content in marine Proteobacteria during non-
steady-state growth. FEMS Microbiol Ecol 30: 253-260.
Kirchman DL (2015). Growth Rates of Microbes in the Oceans. Annu Rev Mar Sci 8: 285-209.
Koop-Jakobsen K, Giblin AE (2010). The effect of increased nitrate loading on nitrate reduction
via denitrification and DNRA in salt marsh sediments. Limnol Oceanogr 55: 789-802.
Langenheder S, Lindström ES, Tranvik LJ (2005). Weak coupling between community
composition and functioning of aquatic bacteria. Limnol Oceanogr 50: 957-967.
Leff JW, Jones SE, Prober SM, et al. (2015). Consistent responses of soil microbial communities
to elevated nutrient inputs in grasslands across the globe. Proc Natl Acad Sci USA 112: 10967-
10972.
11
Lennon JT, Jones SE (2011). Microbial seed banks: the ecological and evolutionary implications
of dormancy. Nat Rev Microbiol 9: 119-130.
Lin X, Feng Y, Zhang H, et al. (2012). Long-term balanced fertilization decreases arbuscular
mycorrhizal fungal diversity in an arable soil in North China revealed by 454 pyrosequencing.
Environ Sci Tech 46: 5764-5771.
Lozupone CA, Knight R (2007). Global patterns in bacterial diversity. Proc Natl Acad Sci USA
104: 11436-11440.
Lynch MD, Neufeld JD (2015). Ecology and exploration of the rare biosphere. Nat Rev
Microbiol 13: 217-229.
Maeda K, Spor A, Edel-Hermann V, et al. (2015). N2O production, a widespread trait in fungi.
Sci Rep 5: 9697.
Mazmanian SK, Round JL, Kasper DL (2008). A microbial symbiosis factor prevents intestinal
inflammatory disease. Nature 453: 620-625.
Mcleod E, Chmura G, Bouillon S, et al. (2011). A blueprint for blue carbon: toward an improved
understanding of the role of vegetated coastal habitats in sequestering CO. Front Ecol Environ 9:
552-560.
Müller AL, de Rezende JR, Hubert CR, et al. (2014). Endospores of thermophilic bacteria as
tracers of microbial dispersal by ocean currents. ISME J 8: 1153-1165.
Pedrós-Alió C (2012). The rare bacterial biosphere. Annu Rev Mar Sci 4: 449-466.
Philippot L, Raaijmakers JM, Lemanceau P, van der Putten WH (2013). Going back to the roots:
the microbial ecology of the rhizosphere. Nat Rev Microbiol 11: 789-799.
Picard KT (2017). Coastal marine habitats harbor novel early-diverging fungal diversity. Fungal
Ecol 25: 1-13.
Raposa KB, Weber RL, Ekberg MC, Ferguson W (2016). Vegetation dynamics in Rhode Island
salt marshes during a period of accelerating sea level rise and extreme sea level events. Estuar
Coasts 15: 1-11.
Reed HE, Martiny JB (2013). Microbial composition affects the functioning of estuarine
sediments. ISME J 7: 868-879.
de Rezende JR, Kjeldsen KU, Hubert CR, et al. (2013). Dispersal of thermophilic
Desulfotomaculum endospores into Baltic Sea sediments over thousands of years. ISME J 7: 72-
84.
12
Rublee P, Dornseif B (1978). Direct counts of bacteria in the sediments of a North Carolina salt
marsh. Estuaries 1: 188–191.
Salazar-Villegas A, Blagodatskaya E, Dukes JS (2016). Changes in the size of the active
microbial pool explain short-term soil respiratory responses to temperature and moisture. Front
Microbiol 7: 524.
Shade A, Peter H, Allison SD, et al. (2012). Fundamentals of microbial community resistance
and resilience. Front Microbiol 3: 417.
Shade A, Jones SE, Caporaso JG, et al. (2014). Conditionally rare taxa disproportionately
contribute to temporal changes in microbial diversity. MBio 5: e01371-14.
Sogin ML, Morrison HG, Huber JA, et al. (2006). Microbial diversity in the deep sea and the
underexplored “rare biosphere”. Proc Natl Acad Sci USA 103: 12115-12120.
Strickland MS, Lauber C, Fierer N, Bradford MA (2009). Testing the functional significance of
microbial community composition. Ecology 90: 441-451.
Tedersoo L, Bahram M, Põlme S, et al. (2014). Global diversity and geography of soil fungi.
Science 346: 1256688.
Thompson RM, Brose U, Dunne JA, et al. (2012). Food webs: reconciling the structure and
function of biodiversity. Trends Ecol Evol 27: 689-697.
Wissuwa J, Stokke R, Fedøy AE, et al. (2016). Isolation and complete genome sequence of the
thermophilic Geobacillus sp. 12AMOR1 from an Arctic deep-sea hydrothermal vent site. Stand
Gen Sci 11: 1.
13
Chapter 1: Nutrient enrichment induces dormancy and decreases diversity of active bacteria in salt marsh sediment
Abstract
Microorganisms control key nitrogen biogeochemical pathways, thus changes in
microbial diversity, community structure, and activity can affect ecosystem function and
response. Factors that control the proportion of active cells in the environment and how that
proportion varies as a function of global change drivers requires clarification. In light of the
increasing supply of anthropogenic nitrogen to ecosystems worldwide understanding how
increasing availability of nutrients alters active microbial communities is imperative. Here we
show that a decade of nitrogen additions to salt marshes, one of the most N enriched ecosystems
in the world, causes a net shift in the active microbial community despite no net change to the
total community. The shift in the active community causes the proportion of dormant microbial
taxa to double, from 45 to 90%, and induces diversity loss in the active portion of the bacterial
community. Our results suggest that perturbations to salt marshes can drastically alter active
microbial communities, however salt marsh microbial communities may remain broadly resilient
by maintaining total diversity through increased dormancy.
Introduction
Human activities have increased the amount of reactive nitrogen (N) in the biosphere.
Excess bioavailable N entering ecosystems can elicit many deleterious effects (Conley et al.
2009) including decreased biodiversity (Isbell et al. 2013, Leff et al. 2015). Numerous studies
have indicated the important coupling between biodiversity and ecosystem function (Hooper et
14
al. 2012), though the emphasis is typically on macro-organismal diversity. Microorganisms,
however, exert strong controls over ecosystems through the mediation of key biogeochemical
cycles, including the N cycle. Despite their importance to ecosystem function, the relationship
between microbial community composition, diversity, and ecosystem function is not clearly
elucidated (Finlay et al. 1997). Resolving the responses of microbial communities, in particular
the active taxa that maintain ecosystem function, is essential for predicting how ecosystems will
respond to global changes such as increased anthropogenic N supply.
Analysis of the 16S rRNA gene, which is commonly used to assess bacterial community
structure, accounts for all cells, including cells that are active, dormant, and recently dead, as
well as extracellular DNA (Nielsen et al. 2007). Relic DNA from dead cells has been shown to
mask ecologically important patterns in soil environments (Carini et al. In press). A large number
of inactive cells might also mask shifts in the active microorganisms that are responsible for
critical ecosystem services. Dormancy is a bet-hedging strategy where microbes enter a low
metabolic or inactive state when they encounter unfavorable environmental conditions (Lennon
and Jones 2011). Microbes can persist in this state until environmental conditions favor their
successful growth. Molecular analyses suggest that dormant microbial taxa can account for ~20-
50% of the microbial community depending on the ecosystem and the heterogeneity of the
environment (Lennon and Jones 2011). It is unclear how high proportions of dormant taxa affect
microbial function, however, dormancy acts as a genomic reservoir allowing for the preservation
of genetic diversity in the presence of unfavorable conditions, and may therefore provide an
important long-term strategy for maintaining ecosystem function (Lennon and Jones 2011).
Coastal nutrient supply, especially by nitrogen (N) in the form of nitrate (NO3-), has
resulted in salt marshes being one of the most nutrient enriched ecosystems in the world, with
15
some systems experiencing nutrient enrichment greater than 500 kg N km-1yr-1 (Pardo et al.
2011). Salt marsh area has significantly declined across the globe and recent evidence suggests
that excess nutrients accelerate marsh collapse (Deegan et al. 2012). Nutrient-induced loss of
marsh area is hypothesized to occur through a decrease in belowground plant biomass and an
increase in belowground bacterial respiration, resulting in decreased stability of the marsh edge
and loss of marsh area. However, evidence from 16S rRNA gene analysis suggests that N-
enrichment has a minimal effect on salt marsh sediment total bacterial community structure
(Bowen et al. 2011). The lack of response in the total bacterial community is surprising given the
observed shifts in biogeochemical function and respiration as a result of nutrient enrichment
(Deegan et al. 2007, Koop-Jakobsen and Giblin 2010) and additional work that clearly connects
changes in ecosystem function with changes in bacterial community structure (Reed and Martiny
2013).
We hypothesize that the apparent lack of bacterial community response to N-enrichment
is a result of a high degree of bacterial dormancy in salt marsh sediments and that the active
community of bacteria will shift in response to increased N supply. Salt marsh sediments are
highly dynamic habitats, where microorganisms are exposed to variable light, oxygen, salinity,
water, carbon, and nutrients that can change in minutes to hours. Because of these widely
changing conditions, we predict that dormant taxa may account for a substantial portion of the
microbial community in marsh sediments. We hypothesize that excess N will favor a small
number of taxa that are able to respond to increased N availability and ultimately result in an
increase in the portion of dormant cells.
Here we test our hypotheses by using analysis of the 16S rRNA gene and the gene
product, 16S rRNA, to assess the response of the total and active microbial communities,
16
respectively, to N-enrichment. Further, we quantified the extent of bacterial dormancy in salt
marsh sediments and the effect of nutrient enrichment on active bacterial diversity. Although the
relative abundance of 16S rRNA cannot be used as an exact proxy for bacterial growth
(Blazewicz et al. 2013) it does indicate cells that have the potential for growth and cells that are
metabolically active, though perhaps not dividing (Kerkhof and Kemp 1999). We examined the
effect of experimental long-term nutrient enrichment (Deegan et al. 2007, 2012, Johnson et al.
2016) on these communities in low and high marsh habitats in a New England salt marsh located
at the Plum Island Ecosystems Long-Term Ecological Research Site. Our long-term
experimental marsh has been enriched with 70 µM nitrate on the incoming tides seasonally since
2004. Another marsh that we sampled was enriched for one year in 2005 and again from 2009-
2015. In each year enriched marshes received 15-fold more N than reference marshes (Deegan et
al. 2007, Johnson et al. 2016).
New England salt marshes are tidal grasslands characterized by low and high marsh
habitats that receive different amounts of tidal flooding. Low marsh is dominated by the tall
ecotype of Spartina alterniflora and is flooded on every semi-diurnal tide. The low marsh habitat
receives considerably more N delivered by tidal water than the high marsh, which is dominated
by Spartina patens and is flooded by only 30% of high tides (Johnson et al. 2016). We collected
triplicate sediment samples from both habitats monthly in each marsh (May-October; 2005,
2006, 2013, 2014) to examine short-term and long-term trajectories in community structure and
activity. After extracting DNA and RNA, we amplified and sequenced the 16S rRNA gene and
16S rRNA to describe the total and active bacterial communities. To examine how changes in the
active microbial community might translate to ecosystem scale processes we also measured
17
whole ecosystem metabolism by examining dissolved oxygen concentrations in creek water of
one enriched and one reference creek.
Methods
Study Site Description
This experiment was conducted at the Plum Island Ecosystems Long-Term Ecological
Research Station (PIE-LTER) in Northeastern Massachusetts, USA (42.759 N, 70.891 W).
Nested within the LTER, a long-term (>10 y), large-scale (60,000 m2 per treatment marsh
system) nutrient enrichment experiment called the TIDE Project (Deegan et al. 2007, 2012,
Johnson et al. 2016) is ongoing. One marsh system received nutrient enrichment with target
enrichment of 70-100 μM nitrate in the creek water on the rising tide continuously throughout
the growing season since 2004 and an additional creek was fertilized in 2005 and then again
from 2009 to the present. Two additional creeks received no fertilizer and serve as reference
marshes.
Sample Collection
We collected samples monthly over the course of the growing season (May-October) in
2005, 2006, 2013, and 2014 at low tide. All samples were collected within 1 m of permanent
transects. Sediment was collected within two dominant macrophytes: Spartina patens (SP) and
the tall ecotype of Spartina alterniflora (TSA). Surface sediments (top 2 cm) were collected in
triplicate with sterile, cut-off 30 mL syringes and homogenized in a sterile 50 mL centrifuge
tube. Samples were aliquoted into either empty cryovials or cryovials containing 1 mL RNAlater
(Invitrogen, Grand Island, NY, USA). Samples put into RNAlater were mixed by vigorous
18
shaking prior to storage. All samples were stored on dry ice for less than 1 hour until they were
flash frozen in liquid nitrogen.
Nucleic Acid Extraction
DNA was extracted from ~0.25 g of sediment using the MoBio® PowerSoil Total DNA
Isolation Kit (Carlsbad, CA, USA) following the manufacturer's instructions. RNA was extracted
from ~0.5 g sediment following the protocol of Mettel et al. (2010) with modifications. First, to
remove residual RNAlater, sediments were spun at 20,000 x g for 1 minute and the resulting
supernatant was discarded. To each sample 700 µL of PBL buffer (5 mM Tris-HCl [pH 5.0], 5
mM Na2EDTA, 0.1% [wt/vol] sodium dodecyl sulfate, and 6% [vol/vol] water-saturated phenol),
along with 0.5 g 0.17 mm glass beads were vortexed at maximum speed for 10 minutes. Samples
were then spun at 20,000 x g for 30 seconds and the supernatant was transferred to a clean tube.
The remaining sediment and glass beads were resuspended in 700 µL TPM buffer (50 mM Tris-
HCl [pH 5.0], 1.7% [wt/vol] polyvinylpyrrolidone, 20 mM MgCl2) and vortexed at maximum
speed for an additional 10 minutes. Sediment was then spun at 20,000 x g for 30 seconds and the
supernatant was pooled with the previous supernatant. An equal volume of
phenol:choloroform:isoamyl alcohol (25:24:1 v/v/v) was added to each sample and mixed by
vortexing at maximum speed for 30 seconds. Samples were then spun at 20,000 x g for 30
seconds, the aqueous layer was transferred to a fresh tube, and nucleic acids were precipitated
with 0.7 volumes of 100% isopropanol and 0.1 volumes 3 M sodium acetate (pH 5.7). Samples
were spun at 20,000 x g for 30 minutes, the supernatant was discarded, and the resulting pellet
was washed with 70% ethanol and allowed to air dry. The washed RNA was loaded onto an
Illustra Autoseq G-50 Spin Column (GE Healthcare, Pittsburgh, PA, USA), which contained 500
19
µl prewashed Q-Sepharose (GE Healthcare). Samples were spun at 650 x g for 7 seconds and
then eluted 5 times with 80 µl of 1.5 M NaCl (pH 5.5). The flow through was transferred to a
clean tube, precipitated with 0.7 volumes 100% isopropanol and 0.1 volumes sodium acetate (pH
5.7), and spun at 20,000 x g for 30 minutes. The resulting pellet was washed with 70% ethanol,
allowed to air dry, and then resuspended in 50 µl di-ethyl pyrocarbonate (DEPC) treated water.
RNA samples were checked for DNA contamination using general bacterial primers
515F and 806R (Caporaso et al. 2011) and any DNA contamination was removed using DNase I
(New England BioLabs, Ipswich, MA, USA) following the manufacturer's instructions. We then
reverse transcribed RNA to cDNA using the Invitrogen Superscript RT III cDNA synthesis kit
following the manufacturer's instructions for random hexamers. Proper cDNA synthesis was
verified by PCR with general bacterial primers, with product checked on a 1.5% agarose gel
stained with ethidium bromide.
PCR and Sequencing
We quantified DNA and RNA concentrations with Picogreen and Ribogreen (Invitrogen)
kits respectively following the manufacturer's instructions and DNA and cDNA were normalized
to 3 ng μl-1 for all PCR reactions. Samples were then prepared for sequencing on the Illumina
MiSeq (Caporaso et al. 2011). We first used general bacterial primers 515F and 806R (Caporaso
et al. 2011), with appropriate Illumina adaptors, and individual 12-bp GoLay barcodes attached
to the reverse primers. We amplified each sample in triplicate using previously described PCR
conditions (Caporaso et al. 2011). Samples were verified with gel electrophoresis and target
bands were excised and purified with the Qiagen QIAquick gel extraction kit (Qiagen, Valencia,
CA, USA). Samples were quantified fluorometrically using a Qubit 2.0 (ThermoFisher,
20
Waltham, MA, USA) and were pooled in equal molar concentrations for sequencing on the
Illumina MiSeq platform for paired-end 151 bp sequencing. All sequencing was performed at the
University of Massachusetts Boston using V2 chemistry.
Sample Preservation and Extraction Method Verification
Due to the nature in which the samples were preserved (RNAlater and dry ice) and how
the nucleic acids were extracted (MoBio PowerSoil DNA Isolation kit and following Mettel et al.
2010), we sought to understand the effect of our preservation and extraction methodology. While
studies have shown RNAlater to be an effective method to preserve rRNA and mRNA for
sequencing (Ottesen et al. 2011), these studies focused on water samples and the efficacy of
RNAlater in preserving sediment samples has not been investigated. To test our methodology,
we extracted three salt marsh sediment samples preserved in either RNAlater or flash frozen in
liquid nitrogen. Additionally, we verified our extraction methods against a commercially
available co-extraction kit (MoBio PowerSoil Total RNA Isolation Kit) following the
manufacturer’s instructions. The 16S rRNA gene and rRNA were sequenced and the resulting
sequence data was processed as described below. To determine any significant differences
between methods we compared weighted UniFrac similarity (Supplemental Fig. 3) values using
a one-way ANOVA. There were no significant differences as a result of the different extraction
or sediment preservation methods (Supplemental Fig. 3). It is also worth noting that the
synchronous patterns in active microbial community structure in the reference marshes over a
decade provide confidence that the long-term storage of sediments in RNALater™ did not
adversely affect the nucleic acids (Fig. 1C).
Sequence and Data Analysis
21
Paired end reads were joined using fastq-join (Aronesty 2011) with default parameters.
Joined reads were then demultiplexed and quality filtered in QIIME (Caporaso et al. 2010)
following methods outlined by previously (Bokulich et al. 2013) Sequences were screened for
chimeras using de novo mode in UCHIME (Edgar et al. 2011) and the resulting chimeric
sequences were discarded. After quality filtering, a total of 25.31 million rRNA and rRNA gene
sequences were included in the final analysis. Operational taxonomic units (OTUs) were
clustered at 97% sequence similarity using swarm (Mahé et al. 2014) in QIIME and OTUs
appearing only once (singletons) across the dataset were discarded. A representative sequence
was chosen from the most abundant sequence for each OTU and taxonomy was assigned using
BLAST against the GreenGenes database (version 13.5). Additionally, we filtered out all
sequences matching chloroplasts. Representative sequences for each OTU were aligned with
PyNast (Caporaso et al. 2010b) and a phylogenetic tree was constructed using fasttree (Price et
al. 2010).
Beta diversity was calculated on normalized OTU tables using weighted UniFrac
(Lozupone and Knight 2005). For comparison, we also analyzed beta diversity using Bray Curtis
similarity, which does not take the phylogenetic information of the OTUs into account when
calculating similarity and the results are consistent with UniFrac results. To assess differences in
community composition, we used Adonis (Anderson 2001) with 10,000 permutations
implemented in QIIME. Adonis tests significance of categorical variables using a permutational
multivariate analysis of variance by fitting linear models to distance matrices and assessing
model fit with an F-test. To assess the degree of dormancy we calculated the 16S rRNA:16S
rRNA gene +1 ratio for each taxa in each sample and defined a taxon as active with a ratio >1
(Jones and Lennon 2010). Work has shown 16S rRNA:16S rRNA gene ratios can vary among
22
taxonomic groups (Blazewicz et al. 2013) and the ratio used here, and previously (Campbell et
al. 2011), may produce many false positives for active taxa due to the variation in the ratio and
biases associated with methodology (Blazewicz et al. 2013, Carini et al. In press). To provide
further support for our dormancy conclusions we addressed this issue by increasing the ratio of
16rRNA to the 16S rRNA gene between 1 and 50 to assess the effects of the ratio on the
calculated rates of dormancy (Supplemental Fig. 2). We tested how levels of dormancy varied
with fertilization using a one-way ANOVA in R (R Core Team 2012).
Shannon Diversity was calculated on rarified OTU tables normalized to the lowest
sampling depth (32,501) in QIIME (10,000 restarts with steps of 100). To assess the diversity of
active taxa, we calculated the 16S rRNA to 16S rRNA gene ratio for each order and defined an
order as active when its ratio was greater than 1. To determine significant changes in diversity
between fertilization regimes and over time we used multiple one-way ANOVAs after testing
that the data met the relevant assumptions. Finally, to assess significant differences in relative
OTU abundance and relative abundance of microbial orders due to fertilization we used a
Kruskal-Wallis test in QIIME and R, defining significance with a Bonferroni corrected p-value
<0.001. Due to the large number of OTUs present in the dataset, we only compared taxa present
100 times when examining differential frequencies of taxa. Due to the conservative nature of
Bonferroni correction, we also examined the results of a false detection method (Benjamini-
Hochberg test) and found that they produced identical results.
Net ecosystem metabolism in tidal creeks
YSI ™ 6 series water property sondes (measuring conductivity, temperature, and oxygen
saturation state) and Onset ® HOBO pressure loggers (for water depth) were deployed in
23
summer of 2012 in the bed of each tidal creek roughly 2 m below the tall-form Spartina
alterniflora zone, at 115 m (N-enriched) and 129 m (reference) upstream from the nearest
confluence with another primary creek. Sonde data were pre- and post-calibrated, and
independently cross-calibrated against in situ oxygen concentration. Concentration is determined
from saturation state using the oxygen solubility function (Garcia and Gordon 1992). Only six
days of sensor data are analyzed here because a bank collapse covered the sensor in the fertilized
creek preventing further data analysis. Water depth and creek geometry were used to construct
tidally varying water balances. Modeled freshwater inputs (Ganju et al. 2012), tidal porewater
exchange (Nuttle and Hemond 1998), and spring-neap driven drainage of the marsh platform
(Gardner and Gaines 2008) are small compared to advective fluxes during the deployment period
(~1% of total flux each) and are not included.
Dissolved oxygen concentrations and the water budget are used to build a non-steady
state oxygen mass balance for each tidal creek (Nidzieko et al. 2014). The oxygen balance
includes terms for gas exchange and the balance between photosynthesis and respiration i.e. net
ecosystem metabolism in oxygen units (NEM), and is solved for a control volume at the
sampling location. We extend the control surfaces into the marsh sediments to include any
oxygen consumption by other terminal metabolic products (e.g. sulfide; Howarth 1984, Larsen et
al. 2015) in the resulting estimate of NEMO2. We chose a gas exchange parameterization that
incorporates short-term variability in both wind speed and current velocity in a shallow, limited
fetch channel (Borges et al. 2004). Gas exchange flux is smaller in magnitude (25-45% lower) if
the current and wind speed or current only parameterizations are used (Borges et al. 2004,
Nidzieko et al. 2014). All four estimates are consistent with in situ noble gas fluxes during the
experimental period at each site.
24
Average and standard errors of NEMO2 (per unit length of creek) over the hour of high
tide at both creeks are plotted in Fig. 5A, when the tall-form Spartina alterniflora zone is flooded
and interacting with the creek water column. Morning high tides may be more heterotrophic in
this particular zone because of oxygen deficit built up overnight or light and temperature
enhanced respiration prior to peak photosynthesis; this trend is consistent with the longer time
series available at the reference creek. The difference between the NEMO2 in the N-enriched and
reference creeks is plotted in Fig 5B. The generally higher daytime and lower nighttime NEMO2
in the N-enriched creek is consistent with both enhanced oxygen production and respiration in
the N-enriched creek.
There are two potential sources of positive bias that may explain NEMO2 values greater
than zero for the three consecutive nighttime spring tides between 23:00 and 01:00: First, we do
not constrain bubble injection (Morris and Whiting 1985, Hamme and Emerson 2006) and the
dissolution of air trapped in the surface sediments as the tide rises; our mass balance would
erroneously attribute oxygen added in this way to biological production. This process is likely to
affect both creeks in a similar way and thus would not affect the NEM difference between the
creeks (which is what is relevant to the arguments presented here). Second, the reference creek
connects to an adjacent drainage during the highest tides (marked with an arrow on Fig. 5A & B)
and therefore the volume transport model likely is not accurate at these highest tides. This factor
would only affect the reference creek. Based on two additional weeks of O2 measurements at the
reference creek only (not shown), we can deduce from the difference of typical night-time high
tide NEMO2 (n=8) and peak flooding NEMO2 (n=3) that this second factor is likely equal to
approximately 0.35 mmol O2 m-1 min-1. Thus the errors caused with drainage during high tide
would decrease the difference between reference and enriched creek but for 2 out of the 3 peak
25
flooding time points (marked with black arrows), the nighttime NEM difference between
reference and enriched creeks would still be significant.
Results
Total and Active Community Composition
Our results revealed a sharp division between the active and total microbial communities,
suggesting the structure of the active community does not reflect the total community (Fig. 1A,
permutational MANOVA; F=19.43, p<0.001). There was no significant effect of N-enrichment
on total community structure (Fig. 1B; p> 0.67) or diversity (Supplementary Fig. 1; p> 0.43),
although we found a significant effect of habitat on total bacterial community structure
(F=19.51, p<0.001). These results extend previous findings that the total salt marsh bacterial
community is resistant to perturbation by nutrient enrichment (Bowen et al. 2011) and indicate
that this resistance has persisted despite a decade of N-enrichment. Additionally, our results
demonstrate a surprisingly low amount of variability in the total microbial community,
suggesting long-term stability of sediment communities despite perturbations, a phenomenon
that was also observed in other New England marshes receiving excess nutrients for over 40
years (Bowen et al. 2011).
Analysis of the potentially active bacterial community shows a markedly different
pattern than the total bacterial community and indicates that long-term nutrient enrichment
spatiotemporally standardizes the active community, overriding the importance of both habitat
and season as structuring forces (Fig. 1C). In reference marshes, habitat (PERMANOVA;
p<0.01, F=9.05) and seasonality (p<0.03, F=24.24) structured active bacterial communities and
these patterns were repeatable over a decade. Active bacterial communities in N-enriched
26
marshes (Fig. 1C), however, were significantly different from reference marshes and were no
longer influenced by habitat-specific factors (Fig. 1C; p<0.001, F=7.75). Further, the seasonal
patterns in the active microbial community were initially present in the early years of fertilization
(Fig. 1C), however, after a decade of fertilization these patterns were no longer present.
Changes in Dormancy and Diversity
To assess the role of dormancy we calculated the proportion of dormant taxa in each
sample using the 16S rRNA:16S rRNA gene ratio and defining any taxon with a ratio ≤ 1 as
dormant (Jones and Lennon 2010). There are caveats to this approach (Blazewicz et al. 2013), so
to ensure our interpretation was robust, we also increased the ratio required from a ratio of one to
50 (Aanderud et al. 2015; Supplemental Fig. 2). Regardless of the ratio threshold we used to
define dormancy (Supplemental Fig. 2), the results show a consistent pattern that the dormant
taxa in N-enriched sediments accounted for a significantly higher proportion of the bacterial
community than in reference sediments (Fig. 2A; ANOVA, p<0.001, F=41.94). Dormant taxa
remained ~45% of the community in reference sediments (Fig. 2A) but in N enriched sediments
the proportion of dormant taxa increased over time (Fig 2A) to ~90% of the community after a
decade of N-enrichment.
The increase in dormancy corresponded to a decrease in potentially active bacterial
diversity. Reference marsh sediments showed no temporal trends in Shannon Diversity and
displayed significantly higher active diversity than N-enriched sediments (Fig. 2B&C; F=12.49,
p<0.01). Furthermore, bacterial communities in N-enriched S. alterniflora displayed a significant
loss in active diversity over time (p<0.001, F=12.43; Fig. 2B). By contrast, although active
diversity in S. patens appears to decrease over time (Fig. 2C), the variation in diversity was
higher and the decrease was not significant. The greater response observed in the low marsh S.
27
alterniflora habitat is likely results from the fact that it is inundated for a longer period of time
than S. patens and therefore receives a greater N supply (Johnson et al. 2016).
Taxonomic Compositional Changes
Due to the observed sharp decline in active bacterial diversity, we assessed which taxa
responded to N-enrichment by examining the ratio of 16S rRNA to the 16S rRNA gene +1 (Fig.
3). Implicit in this calculation is the assumption that if a sequence is present in the 16S rRNA of
a sample it must also be present at least one time on the 16S rRNA gene, but was not detected
due to incomplete sequencing. Three of the five most abundant bacterial orders we identified
(Desulfobacterales, unclassified Cyanobacteria, and Oscillatoriales) had considerably higher
ratios in N-enriched compared to reference marshes. Despite having a few taxa with very high
16S rRNA:16S rRNA gene ratios, N-enriched marshes contained fewer abundant active taxa
(n=23) than nearby reference marshes (n=41). Furthermore, N-enriched marshes contained a
long tail of taxa that had an activity ratio considerably lower than one, indicating that these taxa
were largely dormant in N-enriched marshes but remained active in reference marshes. Our
results suggest, given the large portion of inactive taxa with very low ratios of 16S rRNA: 16S
rRNA gene in N-enriched marshes, that many of these taxa are highly abundant members of the
total community but are inactive due to N-enrichment enhancing the competitive ability of some
taxa at the expense of many others.
Among the most abundant active taxa, defined as those that were present at least 100
times, seven bacterial orders differed significantly (Kruskal-Wallis test, Bonferoni corrected
p<0.001) between N-enriched and reference marshes (Fig. 4, Supplemental Table 1). Retrieved
sequences from N-enriched sediments contained large numbers of anaerobic sulfate reducers
from the Deltaproteobacterial order Desulfobacterales as well as numerous autotrophic
28
Cyanobacteria from the order Oscillatoriales. Whole genome analysis of Desulfobacterium
autotrophicum indicates the order Desulfobacterales may be metabolically diverse due to large
genomes and numerous transposable elements relative to other sulfate reducing bacteria
(Strittmatter et al. 2009), which may allow faster response to environmental perturbations,
including nutrient enrichment. In reference marshes anoxygenic, or potentially anoxygenic
phototrophs (Chromatiales, Chlorobi, Rhizobiales, Rhodocyclales) were significantly more
abundant than in fertilized marshes. By contrast, cyanobacteria dominated the phototrophs in
fertilized marshes (Fig. 3), with unclassified Cyanobacteria and the cyanobacterial order
Oscillatoriales, responding favorably to nutrient enrichment. Oscillatoriales are common bloom-
forming autotrophs that display broad genetic, phenotypic, and habitat diversity, which may
explain their ability to respond to higher N loads despite many members possessing the capacity
to fix N (Marquardt et al. 2007). While N-fixation has been reported in salt marshes (Valiela and
Teal 1979), which are characterized by high concentrations of particle-bound and pore water
ammonium, enhanced N-enrichment to surface sediments may allow Oscillatoria in fertilized
marshes to forego N-fixation and increase their activity relative to Oscillatoria in reference
marshes.
Changes in Ecosystem Metabolism
To determine whether there were any effects on ecosystem function resulting from the
proliferation of Desulfobacterales, unclassified Cyanobacteria, and Oscillatoriales, we measured
whole ecosystem metabolism using dissolved oxygen concentrations in the water column of one
pair of reference and N-enriched marsh tidal creeks (Fig. 5). During high tide, when sediments in
the low marsh habitat were submerged, the N-enriched creek demonstrated both enhanced
autotrophy during the day and enhanced heterotrophy at night compared to the reference creek
29
(Fig. 5). Previous work also suggested that when salt marsh bacterial activity was enhanced by
N-enrichment, it was in response to increased phototrophic productivity (Bowen et al. 2009).
Taken together these data suggest a tight coupling whereby Cyanobacteria in marsh sediments
consume excess nitrate and release labile carbon compounds that stimulate the activity of
Desulfobacterales. Thus, there may be alteration of the carbon cycle in nutrient enriched marshes
due to enhancement of both bacterial photosynthesis and respiration and the role these changes
play in maintaining marsh geomorphology is a critical concern (Deegan et al. 2012).
Discussion
Microbial dormancy in these nutrient enriched salt marsh sediments is higher than
reported values for soil systems (Lennon and Jones 2011), lake systems (Jones and Lennon
2010), and coastal waters (Campbell et al. 2011) using comparable methodology. These results
confirm the assertion that dynamic abiotic conditions may promote dormancy, but also highlight
the important role that dormancy plays in the capacity of microbial communities to withstand
environmental perturbations. Dormancy can be a response to nutrient limitation and there is
evidence that alleviation of nutrient limitation can reduce rates of dormancy in lake systems
(Jones and Lennon 2010). Salt marshes are N limited systems (Nixon et al. 1986) yet despite N-
enrichment, our results indicate that salt marsh microbial communities increase the proportion of
dormant taxa in response to N additions, which may suggest a differential response of water and
soil/sediment communities. We hypothesize that the high rates of dormancy we observed in the
nutrient enriched marshes are a result of enhanced competition whereby bacteria better adapted
to higher N-loads outcompete other bacteria for nutrients and energy sources, leading to an
increase in dormancy among other taxa present in the community. In a sense, nutrient addition to
30
marsh sediments may induce blooms of Desulfobacterales and Oscillatoriales, much like
estuaries often have algal blooms in response to excess N (Anderson et al. 2002). Other nutrient
enrichment induced changes to the marsh, including changes in the geomorphology (Deegan et
al. 2012) of the marsh creeks and changes in faunal abundances (Johnson and Short 2013) could
also influence the proportion of active taxa in marsh sediments. Regardless, our data suggest that
despite the dramatic effect of nutrient enrichment on the active microbial community, the total
bacterial community remained unaffected, allowing marshes to maintain a reservoir of genetic
diversity that provides a buffer such that the highly diverse bacterial community can respond to
future environmental change.
In contrast to terrestrial grasslands where edaphic conditions are comparably more
stable, salt marshes are highly dynamic systems, with environmental conditions changing on
rapid time scales (minutes to hours). Our results demonstrate a surprising stability of the total
microbial community to N-enrichment (Fig. 1B), however, a recent analysis of terrestrial
grasslands systems (Leff et al. 2015) showed consistent shifts in the total microbial community in
response to nutrient enrichment from grasslands across the globe. Our results suggest differential
control mechanisms on the structure of microbial communities between soils and water-saturated
sediments. This difference may be related to community adaptation to intrinsic environmental
variability such that sediment microbial communities respond by altering the dormant versus
active community, thereby allowing them to respond to rapid changes in their environment,
while soil microbial communities alter the total community structure.
Nutrient enriched salt marshes demonstrate a marked loss in active microbial diversity
(Fig. 2b,c), suggesting a net shift in the metabolic functioning of microbial populations, despite
no net change in total community composition. Given the stark decline in active bacterial
31
diversity we expect to see a decrease in functional diversity as well as a truncation of ecosystem
function, as was previously shown in pelagic communities (Gilbert et al. 2010, McCarren et al.
2010). While relationships between biodiversity and ecosystem function have been shown in a
variety of organisms such as plants (Hooper et al. 2012), studies examining the relationship
between diversity and function in microbial communities have been equivocal. Reed and
Martiny (2013) demonstrated a net shift in community composition correlated with enhanced
respiration, however, this correlation was not observed in a similar study (Waldrop and Firestone
2006). Given the decrease in diversity of the active microbial community and the ecosystem-
scale enhancement in both autotrophy and heterotrophy in response to N-enrichment, we
conclude that N-enrichment to salt marshes likely alters metabolic function and thus, ecosystem
function of these critically important coastal habitats.
Increased N additions to ecosystems worldwide have fundamentally altered microbial
communities and the biogeochemical functions they mediate (Galloway et al. 2008, Hooper et al.
2012). In coastal salt marshes, N-enrichment promotes an active bacterial community that is
spatiotemporally homogenous compared to reference marshes. A decade of nutrient enrichment
in marsh sediments resulted in the highest reported proportion of dormant bacterial taxa and
induced diversity loss in the active portion of the bacterial community, despite no apparent
change in the total bacterial community. We suggest that the highly dynamic nature of salt marsh
sediments promotes bacterial dormancy as an ecological strategy for the maintenance of
microbial diversity. Our results suggest a unique feature of salt marsh bacterial communities in
that the active diversity of the community can be dramatically diminished, while still maintaining
a genetic reservoir of traits that can respond to future environmental changes. This work
underscores the importance of examining both total and active bacterial communities in future
32
studies as their responses to global change drivers such as nutrient enrichment are likely to
diverge.
Acknowledgments
would also like to thank researchers of the TIDE Project (NSF OCE0924287, OCE0923689,
DEB0213767, DEB1354494, and OCE 1353140) for maintaining the study site and the N-
enrichment experiment. The TIDE project provided the YSI sondes, Jimmy Nelson did pre- and
post-calibrations on the sondes, and Will Kearney and Sergio Fagherazzi provided the creek
geometry transects for the water budget. We would also like to thank various members of the
Bowen lab for their help in field collections. Thom Thera and Illumina tech support provided
tech support for sequencing. Bioinformatics could not have been done without the use of the
supercomputing facilities managed by Jeff Dusenberry and the Research Computing Department
at the University of Massachusetts Boston. The Plum Island LTER (NSF LTER OCE 0423565
and NSF OCE 1058747) provided critical intellectual and study site support. Finally, we would
like to thank Anne Giblin, David Johnson, James Nelson, Sarah Feinman, Andrew Babbin, Jay
Lennon, and an anonymous reviewer for helpful comments and discussions on the manuscript.
Literature Cited
Aanderud ZT, Jones SE, Fierer N, Lennon JT (2015). Resuscitation of the rare biosphere
contributes to pulses of ecosystem activity. Front Microbiol 6: 24.
Anderson MJ (2001). A new method for non-parametric multivariate analysis of variance.
Austral Ecol 26: 32–46.
Anderson DA, Gilbert PA, Burkholder JM (2002). Harmful algal blooms and eutrophication:
nutrient sources, composition, and consequences. Estuaries 25: 704-726.
33
Aronesty, E. (2011). ea-utils: Command-line tools for processing biological sequencing data,
http://code.google.com/p/ea-utils.
Blazewicz SJ, Barnard RL, Daly RA, Firestone MK (2013). Evaluating rRNA as an indicator of
microbial activity in environmental communities: limitations and uses. ISME J 7: 2061-2068.
Bokulich NA, Subramanian S, Faith JJ, et al. (2013). Quality-filtering vastly improves diversity
estimates from Illumina amplicon sequencing. Nat Meth 10: 57-59.
Borges AV, Vanderborght JP, Schiettecatte LS, et al. (2004). Variability of the gas transfer
velocity of CO2 in a macrotidal estuary (the Scheldt). Estuaries 27: 593-603.
Bowen JL, Crump BC, Deegan LA, Hobbie JE (2009). Increased supply of ambient nitrogen has
minimal effect on salt march bacterial production. Limnol Oceanogr 54: 713-722.
Bowen JL, Ward BB, Morrison HG, et al. (2011). Microbial community composition in
sediments resists perturbation by nutrient enrichment. ISME J 5: 1540-1548.
Campbell BJ, Yu L, Heidelberg JF, Kirchman DL (2011). Activity of abundant and rare bacteria
in a coastal ocean. Proc Natl Acad Sci USA 108: 12776-12781.
Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer N, Pena
AG, Goodrich JK, Gordon JI, Huttley GA (2010). QIIME allows analysis of high-throughput
community sequencing data. Nat Meth 7: 335-336.
Caporaso JG, Bittinger K, Bushman FD, DeSantis TZ, Andersen GL, Knight R (2010b).
PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics 26: 266-
267.
Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, Fierer N,
Knight R (2011). Global patterns of 16S rRNA diversity at a depth of millions of sequences per
sample. Proc Natl Acad Sci USA 108: 4516-4522.
Carini P, Marsden PJ, Leff JW, et al. In press. N. Relic DNA is abundant in soil and obscures
estimates of soil microbial diversity. Nat Microbiol.
Conley DJ, Paerl HW, Howarth RW, et al. (2009). Controlling eutrophication: Nitrogen and
phosphorus. Science 323: 1014-1015.
Deegan LA, Bowen JL, Drake D, et al. (2007). Susceptibility of salt marshes to nutrient
enrichment and predator removal. Ecol Appl 17: S42-S63.
Deegan, L.A, Johnson DS, Warren RS, et al. (2012). Coastal eutrophication as a driver of salt
marsh loss. Nature 490: 388-392.
34
Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R (2011). UCHIME improves sensitivity
and speed of chimera detection. Bioinformatics 27: 2194-2200.
Finlay BJ, Maberly SC, Cooper JI (1997). Microbial diversity and ecosystem function. Oikos 80:
209-213.
Galloway JN, Townsend AR, Erisman JW, et al. (2008). Transformation of the nitrogen cycle:
Recent trends, questions, and potential solutions. Science 320: 889-892.
Ganju NK, Hayn M, Chen SN, et al. (2012). Tidal and groundwater fluxes to a shallow,
microtidal estuary: Constraining inputs through field observations and hydrodynamic modeling.
Estuar Coasts 35: 1285-1298.
Garcia HE, Gordon LI (1992). Oxygen solubility in seawater: Better fitting equations. Limnol
Oceanogr 37: 1307-1312.
Gardner LR, Gaines EF (2008). A method for estimating pore water drainage from marsh soils
using rainfall and well records. Estuar Coast Shelf Sci 79: 51-58.
Gilbert JA, Field D, Swift P, et al. (2010). The taxonomic and functional diversity of microbes at
a temperate coastal site: a ‘multi-omic’ study of seasonal and diel temporal variation. PloS One
5: e15545.
Hamme RC, Emerson SR (2006). Constraining bubble dynamics and mixing with dissolved
gases: Implications for productivity measurements by oxygen mass balance. J Mar Res 64: 73-95
(2006).
Hooper DU, Adair EC, Cardinale BJ, et al. (2012). A global synthesis reveals biodiversity loss as
a major driver of ecosystem change. Nature 486: 105-108.
Howarth RW (1984). The ecological significance of sulfur in the energy dynamics of salt marsh
and coastal marine sediments. Biogeochemistry 1: 5–27.
Isbell F, Reich PB, Tilman D, et al. (2013). Nutrient enrichment, biodiversity loss, and
consequent declines in ecosystem productivity. Proc Natl Acad Sci USA 110: 11911-11916.
Johnson DS, Short MI (2013). Chronic nutrient enrichment increases the density and biomass of
the mudsnail, Nassarius obsoletus. Estuaries and Coasts 36: 28-35.
Johnson DS, Warren RS, Deegan LA, Mozder TJ (2016). Salt marsh plant aboveground
responses to simulated coastal eutrophication. Ecol App. 26: 2649-2661.
Jones SE, Lennon JT (2010). Dormancy contributes to the maintenance of microbial diversity.
Proc Natl Acad Sci USA 107: 5881-5886.
35
Kerkhof L, Kemp P (1999). Small ribosomal RNA content in marine Proteobacteria during non-
steady-state growth. FEMS Microbiol Ecol 30: 253-260.
Koop-Jakobsen K, Giblin AE (2010). The effect of increased nitrate loading on nitrate reduction
via denitrification and DNRA in salt marsh sediments. Limnol Oceanogr 55: 789-802.
Larsen S, Nielsen LP, Schramm A (2015). Cable bacteria associated with long-distance electron
transport in New England salt marsh sediment. Environ Microbiol Rep 7: 175-179.
Leff JW, Jones SE, Prober SM, et al. (2015). Consistent responses of soil microbial communities
to elevated nutrient inputs in grasslands across the globe. Proc Natl Acad Sci USA 112: 10967-
10972.
Lennon JT, Jones SE (2011). Microbial seed banks: the ecological and evolutionary implications
of dormancy. Nat Rev Microbiol 9: 119-130.
Lozupone CA, Knight R (2005). UniFrac: a new phylogenetic method for comparing microbial
communities. Appl Environ Microbiol 71: 8228-8235.
Mahé F, Rognes T, Quince C, de Vargas C, Dunthorn M (2014). Swarm: robust and fast
clustering method for amplicon-based studies. PeerJ 2: e593.
Marquardt J, Palinska KA (2007). Genotypic and phenotypic diversity of cyanobacteria assigned
to the genus Phormidium (Oscillatoriales) from different habitats and geographical sites. Arch
Microbiol 187: 397-413.
McCarren J, Becker JW, Repeta DJ, et al. (2010). Microbial community transcriptomes reveal
microbes and metabolic pathways associated with dissolved organic matter turnover in the sea.
Proc Natl Acad Sci USA 107: 16420-16427.
Mettel C, Kim Y, Shrestha PM, Liesack W (2010). Extraction of mRNA from soil. Appl Environ
Microbiol 76: 5995-6000.
Morris JT, Whiting GJ (1985). Gas advection in sediments of a South Carolina salt marsh. Mar
Ecol Prog Ser 27: 187-194.
Nielsen KM, Johnsen PJ, Bensasson D, Daffonchio D (2007). Release and persistence of
extracellular DNA in the environment. Environ Biosafety Res 6: 37-53.
Nixon SW, Oviatt CA, Frithsen J, Sullivan B (1986). Nutrients and the productivity of estuarine
and coastal marine ecosystems. J Limnol Soc S Afr 12: 43e71.
Nidzieko NJ, Needoba JA, Monismith SG, Johnson KS (2014). Fortnightly tidal modulations
affect net community production in a mesotidal estuary. Estuar. Coasts 37, S91-S110.
36
Nuttle WK, Hemond HF (1998). Salt marsh hydrology: Implications for biogeochemical fluxes
to the atmosphere and estuaries. Glob Biogeochem Cycl 2: 91-114.
Ottesen EA, Marin R, Preston CM, et al. (2011). Metatranscriptomic analysis of autonomously
collected and preserved marine bacterioplankton. ISME J 5: 1881-1895.
Pardo LHM, Fenn ME, Goodale CL, et al. (2011). Effects of nitrogen deposition and empirical
nitrogen critical loads for ecoregions of the United States. Ecol. Appl. 21: 3049-3082.
Price MN, Dehal PS, Arkin AP (2010). FastTree 2–approximately maximum-likelihood trees for
large alignments. PloS One 5: e9490.
R Core Team, R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria, (2012).
Reed HE, Martiny JB (2013). Microbial composition affects the functioning of estuarine
sediments. ISME J 7: 868-879.
Strittmatter AW, Liesegang H, Rabus R, et al. (2009). Genome sequence of Desulfobacterium
autotrophicum HRM2, a marine sulfate reducer oxidizing organic carbon completely to carbon
dioxide. Environ Microbiol 11: 1038-1055.
Valiela I, Teal JM (1979). The nitrogen budget of a salt marsh ecosystem. Nature 280: 652-656.
Waldrop MP, Firestone MK (2006). Response of microbial community composition and function
to soil climate change. Microbiol Ecol 52: 716-724.
37
Figures and Tables
Figure 1. Principal coordinates analysis of weighted UniFrac similarity for (A) the total and
active communities, (B) the total community as assessed by the 16S rRNA gene, and (C) the
active community assessed by 16S rRNA. These data derive from samples collected within one
meter of a permanent transect line that bisected two marsh habitats in four marsh creeks and
were collected monthly from May to October in 2005, 2006, 2013, and 2014 (N=192).
Fert.=fertilized, ref=reference.
38
Figure 2. Percent of dormant taxa in N-enriched (red) and reference (blue) marshes (A) and
Shannon Diversity of the potentially active community, as assed by 16S rRNA for tall Spartina
alterniflora (TSA; B), and Spartina patens (SP; C). Boxes represent 25-75% quartiles, and the
solid black line is the median value.
39
Figure 3. Order-level taxonomic composition of fertilized (red) and reference (blue) sediments
assessed by the log10 16S rRNA:16S rRNA gene ratio for the microbial orders comprising 90%
of all sequences. Orders colored blue are significantly more abundant in reference sediments
(Kruskal-Wallis test, Bonferroni corrected p<0.001) while red orders are significantly more
abundant in N-enriched sediments. Black line is a 16S rRNA:16S rRNA gene ratio of 1. Taxa
with a ratio higher than one are considered active and taxa with ratios lower than one are
considered dormant. Points are the mean and error bars are standard error of the mean.
40
41
Figure 4. Significantly different operational taxonomic units (OTUs) in the potentially active
communities between fertilized (red) and reference (blue) sediments present at least 100 times
assed by a Kruskal-Wallis test (Bonferoni corrected p<0.0001). The class of each OTU is
indicated by the colored circle. More taxonomic information about each OTU can be found in
Supplemental Table 1. Points are the mean and error bars are standard error of the mean.
42
Figure 5. High-tide only dissolved oxygen based net ecosystem metabolism (NEM) in the
fertilized (red) and reference (blue) tidal creeks (A) and the difference (black) between the
fertilized and reference creeks (B) plotted against time of day for each of the 12 high tides during
a six-day period in August 2012. Each point represents the mean NEM or difference rates
evaluated every 10 minutes over the hour of high tide, with one standard error bar for the means.
Note the mean difference rate over each hour of high tide is more tightly constrained than the
mean for either creek. Shading denotes the period between dusk and dawn, and black arrows
identify the time range of three tides that flood the high marsh platform and may have systematic
positive biases in NEM, particularly in the reference creek (see Methods).
43
Chapter 1: Supplementary Information
Supplementary Fig. 1. Box plots of average Shannon Diversity in the total community as assed
by the 16S rRNA gene. Boxes are 25th and 75th quantiles with the medians in bold of samples
collected each year from reference marshes (blue) and fertilized marshes (red). Although the
total diversity of bacteria in S. patens is lower than that of S. alterniflora, there is no difference
in diversity as a result of fertilization, in sharp contrast with what was observed in the diversity
of the active portion of the community.
44
Supplementary Fig. 2. Percent dormant taxa assessed by the ratio of 16S rRNA to the 16S
rRNA gene. Data were generated by increasing the ratio required for a taxon to be considered
active from greater than 1 to 50. At all ratios, fertilized creeks had significantly higher dormant
taxa than reference sediments, as assessed by a one-way ANOVA.
45
Supplementary Fig. 3 Weighted UniFrac similarity comparing community similarity for 16S
rRNA (A) and the 16S rRNA gene (B) for different extraction methods and sample preservation
techniques.
46
Supplementary Table 1- taxonomic information of the 32 significantly different (between
fertilized and reference samples) OTUs present at least 100 times from figure 3 as determined by
BLASTn. Bonferroni corrected p-values are significant at <0.001 as determined by a Kruskal-
Wallis test.
Phylum Class Order Family Bonferroni P
OTU1 Proteobacteria Deltaproteobacteria Desulfobacterales NA 1.25E-14
OTU2 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 6.51E-10
OTU3 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 1.15E-02
OTU4 Cyanobacteria Oscillatoriophycideae Oscillatoriales Phormidiaceae 1.68E-15
OTU5 Proteobacteria Deltaproteobacteria Desulfobacterales Desulfobacteraceae 8.90E-13
OTU6 Proteobacteria Deltaproteobacteria Desulfobacterales Desulfobacteraceae 4.87E-05
OTU7 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 7.80E-11
OTU8 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 3.34E-03
OTU9 Proteobacteria Deltaproteobacteria Desulfobacterales Desulfobacteraceae 1.02E-06
OTU10 Proteobacteria Deltaproteobacteria Desulfobacterales Desulfobacteraceae 1.50E-02
OTU11 Proteobacteria Deltaproteobacteria Desulfobacterales Desulfobacteraceae 2.23E-09
OTU12 Cyanobacteria Oscillatoriophycideae Oscillatoriales Phormidiaceae 2.47E-13
OTU13 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 2.80E-14
OTU14 Proteobacteria Deltaproteobacteria Desulfobacterales Desulfobacteraceae 4.26E-09
OTU15 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 4.14E-10
OTU16 Proteobacteria Deltaproteobacteria Desulfarculales Desulfarculaceae 7.03E-11
OTU17 Proteobacteria Alphaproteobacteria Rhodospirillales Rhodospirillaceae 1.67E-09
OTU18 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 1.28E-02
OTU19 Cyanobacteria Oscillatoriophycideae Oscillatoriales NA 7.71E-14
OTU20 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae 2.75E-06
OTU21 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae 5.19E-09
OTU22 Proteobacteria Alphaproteobacteria Rhizobiales Hyphomicrobiaceae 1.21E-08
47
OTU23 Proteobacteria Alphaproteobacteria Rhodospirillales Rhodospirillaceae 2.11E-05
OTU24 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae 2.40E-06
OTU25 Proteobacteria Deltaproteobacteria Desulfarculales Desulfarculaceae 1.73E-15
OTU26 Proteobacteria Alphaproteobacteria Rhizobiales Hyphomicrobiaceae 8.47E-06
OTU27 Planctomycetes Planctomycetia Pirellulales Pirellulaceae 7.47E-19
OTU28 Proteobacteria Deltaproteobacteria Desulfarculales Desulfarculaceae 8.93E-17
OTU29 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae 3.46E-12
OTU30 Proteobacteria Deltaproteobacteria Desulfarculales Desulfarculaceae 7.18E-14
OTU31 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae 7.79E-03
OTU32 Proteobacteria Deltaproteobacteria Desulfarculales Desulfarculaceae 1.26E-02
48
Chapter 2: Nutrient enrichment alters salt marsh fungal communities and increases
decomposition of salt marsh sediments
Abstract:
Enrichment of ecosystems with excess nutrients, including nitrogen, is occurring at an alarming
rate and has fundamentally altered ecosystems worldwide. Salt marshes, which lie at the land/sea
interface, are highly effective at removing anthropogenic nutrients through the action of
macrophytes and through microbial processes in coastal sediments. The response of salt marsh
bacteria to excess nitrogen has been documented, however, the role of fungi and their response to
excess nitrogen is poorly understood. Here, we document the response of salt marsh fungal
communities to long-term excess nitrate in four distinct marsh habitats within the marsh complex
at the Plum Island Ecosystems Long-Term Ecological Research site in northeastern
Massachusetts, USA. We show that salt marsh fungal communities varied as a function of salt
marsh habitat, with both fungal abundance and diversity increasing with carbon quantity.
Nutrient enrichment significantly altered fungal communities in all habitats through an increase
in fungal abundance and, notably, the proliferation of taxa that have the capacity for
denitrification. Nutrient enrichment also appears to promote decomposition of marsh peat in low
marsh surface sediments, which suggests that fungi, in addition to bacteria, play an important
role in anaerobic decomposition of marsh organic matter, with consequences for long-term
carbon storage.
Introduction:
Enrichment of coastal waters by excess nutrients, in particular nitrogen (N) in the form of
nitrate (NO3-), is occurring at an accelerated pace and is altering global biogeochemical cycles
(Galloway et al., 2008). The addition of excess N to coastal systems, which are typically N
limited (Nixon et al., 1986), can have numerous deleterious effects including increased harmful
algal blooms (Paerl, 1997), decreased species diversity (Elahi et al., 2015), altered
biogeochemical cycling (Koop-Jakobsen & Giblin, 2010), and expanded areas of hypoxia or
anoxia (Rabalais et al., 2002). Salt marshes, which are coastal grassland ecosystems, are highly
49
effective at removing excess nitrogen from coastal areas through trapping nitrogenous
compounds in sediments (Brin et al., 2010), macrophyte uptake (Valiela et al., 1973), and
through denitrification, the microbially-mediated conversion of NO3- to a gaseous end product
(Valiela et al., 1973, Hopkinson & Giblin 2008, Zumft, 1997). Relative to burial and uptake by
macrophytes, denitrification can account for a substantial portion of fixed N loss from marsh
sediments (Valiela and Teal, 1979) and, at times, can account for the majority of nitrogen loss
from marsh ecosystems (Howes et al., 1996).
In addition to removing anthropogenic nutrients, salt marshes are also highly effective at
storing carbon (McLeod et al. 2011). The sequestration of carbon by salt marshes and other
coastal systems, termed “Blue Carbon”, is orders of magnitude greater per hectare than
sequestration in terrestrial systems due to their high productivity, ability to trap organic material,
and anoxic and sulfidic sediments that impede decomposition (Mcleod et al., 2011). The ability
of salt marshes to store carbon for extended periods of time is a balance between the amount of
carbon they fix and the amount of carbon that is removed through decomposition and export.
Therefore, factors such as water supply, nutrient availability, and the redox status of the
sediments, all influence whether salt marshes are net carbon sources or sinks (Kirwan and Mudd,
2012). Multiple human drivers, including nutrient enrichment, can alter the forces that determine
the strength of marshes as a carbon sink. Increased supply of anthropogenic nutrients can
stimulate primary productivity of nutrient limited macrophytes (Fox et al., 2012), decrease
belowground macrophyte biomass (Deegan et al., 2012), stimulate decomposition of existing
organic matter (Deegan. et al., 2012), alter biogeochemical cycling (Hammersley & Howes,
2005; Koop-Jakobsen & Giblin 2010; Peng et al., 2016), and alter the structure of the active
microbial community (Kearns et al., 2016). Thus, understanding the ecosystem-wide response to
excess nutrients is important for predicting how the carbon carrying capacity of marshes will
change by future nutrient enrichment.
The activity of salt marsh bacteria and the controls over their decomposition processes is
well understood (Hammersley & Howes, 2005; Koop-Jakobsen & Giblin, 2010; Bowen et al.
2011; Peng et al., 2016; Kearns et al. 2016), however, we lack a solid understanding of salt
marsh fungi and the factors that influence their composition and activity with respect to carbon
cycling. Fungi are among the most diverse groups of eukaryotic organisms on the planet with a
50
projected five million plus species globally (O'Brien et al., 2005; Blackwell, 2011; Taylor et al.,
2014). Fungal taxa in terrestrial systems are well characterized and their roles as decomposers of
organic matter, plant and animal pathogens and parasites, and symbionts are well documented
(Kirk et al., 2008). Of the described ~100,000 fungal taxa (Kirk et al., 2008), however, less than
one percent have been characterized from marine environments (Kis-Papo, 2005; Richards et al.,
2012). Work on marine fungi has focused on deep ocean sediments (Nagano et al., 2010),
hydrothermal vents (Burgaud et al., 2015), oxygen minimum zones (Stoeck et al., 2006), and
planktonic marine fungi (Wang et al., 2014). In the coastal zone fungi, primarily saprotrophs,
appear to be structured by carbon supply (Dini-Andreote et al., 2016), but changes in salinity,
nutrient availability, and temperature are also important determinants of marine fungal
community structure (Taylor & Cinliffe, 2016). Since fungi are critical to carbon cycling and
because they appear to be abundant in coastal systems (Dini-Andreote et al., 2016; Taylor &
Cunliffe, 2016), it is imperative that we understand how coastal marine fungi respond to global
change drivers, and ultimately, how that translates into changes in the carbon storage capacity of
coastal habitats.
Here, we studied two salt marsh creeks experimentally enriched with nutrients (Deegan et
al., 2007) and two near-by reference creeks that received no excess nutrient supply. In each
marsh we collected samples from four habitats within each site to understand the distribution of
fungi throughout the marsh and we also examined the effect of nutrient loading on fungal
community composition and diversity by comparing results from nutrient enriched and reference
marshes. We hypothesized that salt marsh fungal community structure and abundance would
vary as a function of habitat due to differences in abiotic conditions. Further, we hypothesized
that, like bacterial community composition (Bowen et al., 2009, 2011; Kearns et al., 2016), salt
marsh fungal community composition would be resistant to perturbation by increased nutrient
supply. Our results demonstrate marsh fungal communities do vary as a function of habitat,
however, unlike what we observed with bacterial communities, nutrient enrichment played an
important role in structuring fungal community composition, particularly in habitats receiving
the highest N exposure. Further, excess nitrogen stimulated the decomposition of organic matter
and lowered organic matter quality, with important consequences for salt marsh carbon storage.
51
Methods
Study site description and Sample Collection
Our experiment was carried out at the Plum Island Ecosystems Long-term Ecological
Research Site (PIE-LTER) located in Northeastern Massachusetts, USA (42.759 N, 70.891 W).
Within the LTER, a long-term nutrient enrichment experiment called the TIDE project has been
ongoing for 13 years (Deegan et al., 2007). One marsh creek has received nutrient enrichment in
the form of NO3- dissolved in creek water on the rising tide continuously throughout the growing
season (May-September) since 2004. A second marsh creek received nutrients in 2005 and then
again from 2009 to the present. Two additional creeks, which have similar biophysical properties
as the nutrient enriched creeks, receive no excess nutrients and serve as reference creeks.
Nutrient enriched creeks receive approximately 15 times more N than reference creeks (Deegan
et al., 2007).
We collected samples in May, July, and September of 2013 and 2014 at low tide within
one meter of permanent transects at all four marsh creeks. We collected sediment from four
distinct marsh habitats (SI Fig. 1): mudflat (MF), tall Spartina alterniflora (TSA), S. patens (SP),
and short S. alterniflora (SSA). The four marsh habitats sampled represent a wide range of
abiotic conditions due to their different elevations above mean sea level. The mudflat at tall S.
alterniflora habitats, referred to as the low marsh, are lower in elevation and have longer
inundation times. The S. patens and short S. alterniflora habitats, referred to as the high marsh,
are higher in elevation and are only inundated on the highest high tides. As a result of these
differences in elevation, these habitats have distinct differences in salinity, nutrient availability,
carbon content, pH, redox conditions, soil temperature, soil water content, and plant above and
belowground biomass. At each sampling time and location we collected surface sediments (top
2-cm) in triplicate with sterile cut-off 30-cc syringes, homogenized the sediments in a sterile 50-
mL centrifuge tube, and aliquoted them into cryovials for storage on dry ice. The remaining
sediment in the 50-mL centrifuge tube was stored on dry ice for sediment organic matter
analyses. All samples for molecular analyses were kept at -80°C until extraction and all nutrient
analysis samples were kept at -20°C until processing.
52
Nucleic acid extraction, PCR, and sequencing
DNA was extracted from approximately 0.25 g of sediment using the MoBio PowerSoil®
Total DNA Isolation Kit (Carlsbad, CA, USA) following the manufacturer’s instructions. DNA
extracts were verified with gel electrophoresis. Fungal communities were amplified in triplicate
with primers ITS1F and ITS2 (Walters et al., 2016) targeting the first fungal internal transcribed
spacer region (ITS). Primer constructs were similar to those described previously (Kozich et al.,
2013) and contained overhang sequences to allow for downstream addition of dual indexes and
Illumina adapters. PCR was performed in triplicate 25 μL reactions with 12.5 μL Phusion Hot
Start Master Mix (New England Biolabs, Ipswich, MA, USA), 0.25 μL of each primer (20 pmol
μL-1), 1-4 ng of DNA, and 11 μL of PCR grade water. Samples were amplified with the
following cycling conditions: 94°C for 3m followed by 35 cycles of 94°C for 30s, 50°C for 60s,
72°C for 90s, and a final extension at 72°C for 10m. PCR reactions were checked with gel
electrophoresis and samples were purified with the Qiagen QiaQuick Gel Purification Kit
(Valencia, CA, USA). A second 8-cycle PCR was performed with the Illumina Nextera XT2 kit
following the manufacturer’s instructions to ligate dual indicies and Illumina adaptors for each
sample. Following the second PCR, samples were purified with a Qiagen PCR purification kit,
quantified with a Qubit fluorometer (Thermofisher, Waltham, MA, USA), and pooled in equal
molar concentrations for paired-end sequencing on an Illumina MiSeq. All sequencing was
performed at the University of Massachusetts Boston using V2 chemistry.
Fungal abundance
To determine fungal abundance, we performed quantitative PCR (qPCR) using the primer
pair ITS1F and ITS2 (Walters et al., 2016). DNA and standards prepared from purified ITS PCR
product were quantified with a Qubit fluorometer and serially diluted (Thermofisher) and all
samples were normalized to 3 ng µL-1. Each sample was amplified in triplicate, along with
standards and internal controls on a Stratagene MX-3005p quantitative thermocycler (Stratagene,
La Jolla, CA, USA). ITS genes were amplified in triplicate 25 µL reactions using 0.25 µL of
each primer, 12.5 µL of Qiagen QuantiTect SYBR Green PCR Master Mix, 1 µL of DNA
template, and 11 µL of PCR grade water using conditions described above. Proper product
53
formation was verified with melt curves and gel electrophoresis. All standard curves possessed a
high degree of linearity (R2 > 0.99) and qPCR efficiency ranged from 95-101%.
Elemental analyses and Fourier Transform Infrared Spectroscopy
Elemental composition (percent nitrogen and carbon) was measured on sediments dried
at 50°C for three days on a Perkin Elmer 2400 Series Elemental Analyzer following standard
procedures. We also used Fourier Transform Infrared Spectroscopy (FTIR) to characterize
organic matter quality and to indicate the difference in extent of decomposition as a result of
nutrient enrichment. This technique provides rapid, detailed information about the relative
abundance of organic matter functional groups. To prepare samples for FTIR analysis, we oven-
dried (50ºC) and finely ground the sediments using a mortar and pestle. We analyzed the samples
in diffuse reflectance mode on a Bruker Vertex 70 FTIR spectrometer (Bruker Optics Inc.,
Billerica, MA) outfitted with a Pike Auto Diffuser Accessory (Pike Technologies, Madison, WI)
and obtained the data as pseudo-absorbance (log[1/Reflectance]). Since we were interested in the
MidIR range, we collected at a 2 cm-1 resolution, with 64 co-added scans per spectrum, from
4000 to 400 cm-1 using potassium bromide (KBr) for background correction (Essington, 2004).
We imported the resulting spectra using Unscrambler X (Camo Software (version 10.1),
Woodbridge, NJ) and corrected the data using a calculated two-point linear tangential baseline.
To reduce the dimensionality of the data and compare spectra across site and habitat we
visualized the relationship among carbon species with a correspondence analysis using the vegan
package (Oksanen et al., 2012) in R (R Core Team, 2012). To better understand the difference in
carbon due to habitat and nutrient enrichment, we calculated the ratio of hydroxylamines
(wavelength 3402 cm-1, Table 1) to multiple wavelengths of aromatic carbon (wavelengths
1650/920/840 cm-1; Table 1; Parikh et al., 2014). Higher hydroxylamine:aromatic carbon ratios
indicate more labile carbon. To estimate the extent of decomposition in each habitat and as a
result of nutrient enrichment, we calculated Index I (Veum et al., 2014), which takes into account
seven distinct wavelengths (Table 1). The numerator corresponds to wavelengths associated with
aromatic carbon and the denominator corresponds to wavelengths associated with aliphatic
carbon and each number represents the spectra values measured at the wavelength identified in
the equation:
54
����� � = 1650 + 920 + 840
2924 + 2850 + 1470 + 1405
Index I assesses decomposition because as organic matter is decomposed, it accumulates less
labile aromatic carbon through decomposition of labile aliphatic carbon. Thus, as decomposition
increases, the value for aromatic carbon goes up and the value for aliphatic carbon goes down,
thus increasing the value of Index I (Chefetz et al., 1998; Hsu & Lo, 1999). Increases in Index I
between reference and nutrient enriched marshes would indicate greater decomposition of
organic matter. We assessed significant differences in the ratio of hydroxylamine to aromatic
carbon and in Index I with a repeated measures ANOVA and multiple comparisons were
corrected with a Tukey HSD test.
Sequence processing and analysis
A total of 1.37 million paired-end reads were joined in QIIME (version 1.9.1; Caporaso
et al., 2010) with fastq-join (Aronsety, 2011) using default parameters. Sequences were then
demultiplexed and quality filtered following Bokulich et al., (2013). Reads were checked for
chimeras using USEARCH (Edgar, 2010) in reference and de novo modes and all chimeric
sequences were discarded. To improve ITS analyses, we removed ribosomal fragments from our
sequences using itsX (Bengtsson-Palme et al., 2013). Following quality filtering, 1.2 million
reads remained. Operational taxonomic units (OTUs) were picked against the UNITE database
(version 7; Abarenkov et al., 2010) at 97% sequence identity (Walters et al., 2016) using
UCLUST (Edgar, 2010). OTUs appearing only once and OTUs matching protists were discarded
from the dataset.
We calculated beta diversity on normalized OTU tables (8,000 sequences per sample)
using Bray-Curtis similarity and results were visualized with a principal coordinates analysis
(PCoA). Significant differences in community composition among habitats and between nutrient
enrichment regimes were tested with a non-parametric permutational MANOVA (Anderson,
2002) with 10,000 permutations. To further assess the effect of nutrient enrichment on different
habitats, we compared Bray-Curtis similarity values between reference and nutrient enriched
55
marsh creeks and assessed significance with an ANOVA in R (R Core Team 2012), correcting
for multiple comparisons with a Tukey HSD test. Alpha diversity was calculated on a rarefied
OTU table (10,000 restarts, steps of 100) normalized to the lowest sequencing depth. We
calculated Shannon Diversity and assessed significant differences with an ANOVA in R. To
assess the role of carbon content on fungal community structure, diversity, and abundance we
used an analysis of covariance (ANCOVA) including habitat as random effect and fertilization as
a fixed effect. Using an ANOVA we checked that the covariate (habitat) and the resulting slope
was linear and did not significantly differ from the independent variable’s slope. Further, the
covariate was independent of fertilization. To assess which OTUs were responding to nitrogen
additions we calculated significant differences in OTU frequencies between reference and
nutrient enriched marshes with a Kruskal-Wallis test and defined taxa as significantly different
with a Benjamini-Hochberg corrected p-value <0.01. Finally, we compared our fungal dataset to
previously described fungal denitrifiers from Meada et al., (2015) to determine the extent of
fungal denitrifiers in our dataset.
Results
Community composition, diversity, and abundance
A principal coordinates analysis (Fig. 1A) of Bray-Curtis similarity revealed a significant
effect of both habitat (PERMANOVA, p<0.001, F(3,92)=29.43) and nutrient enrichment (p<0.001,
F(1,94)=31.42) on fungal community composition. Fungal communities associated with all four
habitats experienced significant shifts in community composition as a result of nutrient
enrichment, however, habitats in the low marsh that receive the highest nitrogen exposure
(mudflat and tall S. alternifora) experienced the greatest effect, as indicated by the low similarity
ratio between nutrient enriched and reference marshes (Fig. 1B). Diversity of fungi, as measured
by the Shannon Diversity Index, ranged from two in in reference mudflat habitats to upwards of
four in high marsh habitats (Fig. 2A). Shannon Diversity in nutrient enriched creeks was
significantly higher than in reference creeks in all four habitats (ANOVA, p<0.001,
F(1,94)=29.12). Quantification of fungal abundance using qPCR displayed a similar pattern to
Shannon Diversity (Fig. 2B). The abundance of fungi in reference creeks ranged from ~105 ITS
copies g-1 sediment in the mudflat to upwards of 108 ITS copies g-1 sediment in the high marsh
56
sediments (SP and SSA, Fig. 2B). Nutrient enrichment, however, significantly increased fungal
abundance (p<0.001, F(1,94)=34.56) in all habitats except for the mudflat, which was also very
close to the significance threshold (p=0.07, F(1,23)=19.1).
Taxonomic composition
Salt marsh sediments were dominated by fungal taxa from the phylum Ascomycota
(~84%) with a small presence from the phyla Basidiomycota (~5%) and Chytridiomycota (~2%;
Fig. 3). Additionally, 8% of fungal taxa could not be assigned to a known fungal phylum. All
habitats were dominated by a consistent presence of several families including
Teratospheariaceae, Mycosphaerellaceae, Physalacriaceae, Lasiosphaeriaceae, and from the
orders Capnodiales and Rhytismatales. Taxonomic composition in all habitats shifted in the
nutrient enriched marshes (Fig. 3), consistent with the beta diversity analyses (Fig. 1).
Taxonomic response to NO3- addition and enhancement of fungal denitrifiers
We calculated significant differences in abundance of OTUs between fertilized and
reference marshes using a Kruskal-Wallis test (Fig. 4). Our analysis revealed 23 OTUs with
significantly different abundances between nutrient enriched and reference marshes, with the
majority (n=16) more abundant in nutrient enriched marshes. Many taxa whose abundance
increased with nutrient enrichment, such as taxa from the order Hypocreales and class
Sordariomycetes, were in low abundance in reference creeks. Further, many of the taxa that
increased in abundance in the nutrient enriched creeks were closely related to recently identified
fungal denitrifiers (Meada et al., 2015). We screened our dataset against ITS data from cultured
fungi that are known to denitrify (Fig. 5; Meada et al. 2015). The percent of fungal denitrifiers
was very nearly zero in most reference marsh habitats, but was significantly higher (ANOVA,
p<0.001, F(3,43)=24.01) in reference TSA sediments. Nutrient enrichment significantly increased
relative abundance of fungal denitrifiers in all habitats (p<0.001, F(1,94)=29.43).
Environmental correlates of composition, abundance, and diversity
We used an analysis of covariance (ANCOVA) to examine how fungal composition,
diversity, and abundance varied as a result of carbon content in the sediment, a factor that
essentially scales with habitat (Fig. 6). ANCOVA of the first principal coordinate values from
the PCoA (Fig. 1A) as a function of the percent carbon in the sediment revealed a strong
correlation (R2=0.88) in reference marshes. Axis 1 values in nutrient enriched marshes, however,
were less strongly correlated with carbon (R2=0.56; Fig. 6A). Regressions in both fertilized
57
(F=19.53, p=0.01) and reference (F=96.43, p<0.01) marshes were significant, however, the
slopes of the lines were significantly different (F=9.23, p=0.003). ANCOVA of percent carbon
against Shannon Diversity (Fig. 6B) in reference (R2=0.67) and fertilized (R2=0.09) marshes
indicated a tight and significant (p<0.01) correlation between carbon and fungal diversity in
reference (p<0.01) and fertilized (p=0.02) marshes. Further, the slopes were significantly
different between fertilized and reference marshes (F=41.43, p<0.001). Fungal abundance (Fig.
6C) significantly (p<0.01) correlated (R2=0.90) with carbon in reference marshes, however, this
correlation while still significant (p=0.02) was not as strong in fertilized marshes (R2=0.46).
Additionally, regressions between fertilized and references marshes were significantly different
(F=31.43, p<0.01)
Fourier Transform Infrared Spectroscopy (FTIR)
Fourier Transform Infrared Spectroscopy (FTIR) was used to assess changes in organic
matter quality. A correspondence analysis derived from the full spectra of data from all four
habitats (SI Fig. 2) revealed a large separation (Fig. 7A; PERMANOVA, F(1,47)=24.46, p<0.01)
between nutrient enriched and reference samples in the low marsh (mudflat and tall S.
alterniflora), but no such effect was observed for the high marsh S. patens and short S.
alterniflora sediments. Index I (Fig. 7B), a metric signifying the extent of carbon decomposition,
indicated that nutrient enriched mudflat (F(2,22)=29.01, p<0.01) and tall S. alterniflora sediments
(F(2,22)=34.01, p<0.001) displayed significantly more decomposition than reference sediments.
The ratio of hydroxylamines to aromatic compounds (Fig. 7C), a ratio that increases with the
increased lability of carbon, indicated that the lability of carbon was significantly lower in
nutrient enriched low marsh habitats (F(2,46)=20.44, p<0.01), but no effect on carbon lability in
the high marsh.
Discussion
Salt marshes provide critical ecosystem services to coastal communities, including long-
term storage of carbon. The ability of salt marshes to store carbon, however, is subject to
anthropogenic threats including the supply of excess nutrients. Anthropogenic changes to salt
marshes can have profound effects on the microorganisms that ultimately control carbon
decomposition and storage. In this study, we investigated the composition, diversity, and
58
abundance of salt marsh fungal communities and their response to excess nitrogen in four
distinct habitats in a New England salt marsh (SI Fig. 1). Fungal communities in reference
marshes varied across salt marsh habitats (Fig. 1A, Fig. 3), reflecting the differences in abiotic
conditions that result from subtle changes in marsh elevation among habitats. We identified (Fig.
3) several fungal groups, primarily from the orders Capnodiales, Rhytismatales, Sordariales, and
Myriangiales, that were present across all habitats, suggesting a tolerance to a wide range of
abiotic conditions. Further, recent work in coastal sediments (Picard, 2017), marine planktonic
communities (Taylor & Conliffe, 2016), and a European salt marsh (Dini-Andreote et al., 2016)
documented the importance of these groups, suggesting they are key fungal taxa in coastal
ecosystems. Members of these orders are predominantly saprotrophs (Kirk et al., 2008) and are
likely decomposing portions of the large standing stock of living and dead Spartina that
dominate salt marshes (Valiela et al., 1976; Benner et al., 1984; Johnson et al., 2016).
Our study documented a similar, yet distinct response of salt marsh fungi to fertilization
and changes in abiotic conditions than terrestrial fungi. Salt marsh fungal communities were
strongly structured by habitat and the associated abiotic conditions, supporting previous work in
marshes (Dini-Andreote et al., 2016). Soil fungal communities have been shown to be structured
at broad (Leff et al. 2015) and narrow geographical scales by changes in soil acidity, climate,
plant community composition, organic matter content, and nutrient availability. The addition of
excess N, however, significantly altered community composition of marsh fungi; a response
consistently observed in soil fungal communities (Treseder, 2004; Allison et al., 2008; Lin et al.,
2012; Leff et al. 2015; Zhou et al., 2016) and their associated activity (Allison et al., 2008). A
common change in soil fungal communities in response to excess N is the loss of endophytic
fungi due to the decreased need of the plant host in a nutrient rich environment. Of the taxa that
responded to N-enrichment, only one taxa from the genus Lulwona (Fig 4), which responded
negatively to N additions, are known to be endophytes (Torta et al., 2015). The lack of a
wholescale response of endophytic fungi in marsh sediments may be due to the mild response of
marsh plants to N additions (Johnson et al., 2016) relative to terrestrial plants (Wei et al., 2013;
Borer et al., 2014) or perhaps a primer bias limiting our ability to detect endophytic lineages or
our lack of knowledge about marine endophytic fungi. In addition to changes in community
composition and diversity, we observed increased decomposition (Deegan et al., 2012) and
59
accelerated carbon cycling (Kearns et al., 2016) in response to excess N in marsh systems. By
contrast, several studies of soil fungi have demonstrated depressed rates of fungal decomposition
under excess N (van Diepen et al., 2017) likely as a result of a shift away from lignin
decomposition to cellulose decomposition and perhaps interactions with bacteria. The
discrepancy between decomposition under excess N between underscores a differential response
between soil and marine communities with important implications for carbon storage in marine
environments.
Long-term exposure to excess NO3- significantly altered the composition of fungi across
all habitats, resulting in more diverse and abundant fungal communities (Fig. 2). Members of the
class Sordariomycetes, in particular from the order Hypocreales, which were in low abundance
(<1%) in references marshes, responded the most dramatically to excess NO3- (Fig. 4 and 5).
Their proliferation in nutrient enriched marshes suggests these taxa are better able to use excess
NO3-, either through assimilation into biomass or in dissimilatory processes as an electron
acceptor under anaerobic conditions. Several taxa from the class Dothideomycetes, in particular
from the order Capnodiales, significantly decreased their abundance coincident with excess NO3-
. The decrease in their abundance suggests they are likely outcompeted by taxa better adapted to
higher nitrogen loads, as previously shown for active salt marsh bacteria (Kearns et al., 2016)
and macrophytes (Levine et al., 1998; Pennings et al., 2005; Fox et al., 2012). Further, nutrient
addition can also modulate genomic content and size of microorganisms (Leff et al. 2015) which
may provide adaptations under copiotrophic environments such as those observed in nutrient
enriched creeks. While nutrient enrichment can directly affect the activity of fungi, it likely also
indirectly affects fungi through alterations to salt marsh primary producers (cyanobacteria,
benthic microalgae, macrophytes) and changes in soil chemistry (e.g. pH) highlighting the
response of fungal communities in marsh sediments is likely a multifaceted process.
Decomposition of organic matter is typically slow in oxygen-depleted salt marsh
sediments because the dominant anaerobic metabolisms (sulfate reduction and fermentation)
decompose marsh carbon at a slower rate than in oxygenated soils (Kristensen and Ahmed,
1995). The addition of NO3-, a more energetically favorable electron acceptor, stimulates
denitrifying bacteria (Koop Jakobsen & Giblin 2010) and fungi that can decompose less labile
carbon that was inaccessible to sulfate reducing and fermenting bacteria (Fig. 7). Decomposition
60
by denitrifiers likely alters the forms of carbon available (Solomon et al., 2007) and opens more
niche space for additional fungal taxa (Ekschmitt et al, 2008), thereby increasing overall fungal
diversity (Fig. 2). The increase in fungal denitrifiers (Fig. 5) and the alteration of organic matter
was most pronounced in the low marsh (mudflat and tall S. alterniflora), which receives excess
NO3- on every flooding tide. The high marsh habitats receive excess NO3
- only when flooded
(~30% of the time; Johnson et al., 2016) and the response in these habitats was more muted in
than in the low marsh habitats, implying that both the form (oxidized versus reduced nitrogen)
and the duration of exposure play important roles in determining how anthropogenic nutrient
supply alters the organic matter quality. The overall amount of organic matter did not
significantly vary between fertilized and reference marshes within a habitat, suggesting enhanced
decomposition did not alter the overall amount of carbon in surface sediments (top 2cm).
However, the Plum Island marsh accretes sediment at a rate of ~3mm yr-1 (Fagherazzi et al.,
2012) suggesting input of organic matter either through new sediment or by plant deposition
(Fox et al., 2012) may offset any direct effects of enhanced decomposition we can observe in
surface sediments. Further work is needed to characterize sediments deeper than 2 cm, where the
majority of carbon is stored in marshes, and determine how excess nitrate effects deep, long-term
carbon storage.
Our results demonstrate that excess N stimulates fungal denitrifiers (Fig. 4 and 5).
Denitrification is typically the dominant N-removal pathway in coastal systems (Hopkinson &
Giblin, 2008; Brin et al., 2010) and rates of denitrification increase under excess nitrogen
(Hammersely & Howes, 2005; Koop-Jakobsen & Giblin, 2010; Peng et al., 2016).
Denitrification is an important ecosystem service, but incomplete denitrification can result in the
release in nitrous oxide (N2O), a potent greenhouse gas (Ravishankara et al., 2009). Currently, all
known fungal denitrifiers lack the nosZ gene (Meada et al., 2015) which catalyzes the reduction
of N2O to N2 (Zumft, 1997). Stimulating fungal denitrifiers (Fig. 4 and 5) through excess NO3-
supply may therefore result in a concomitant increase in N2O production. Decreased nosZ
abundance in N-enriched marshes (Kearns et al., 2015) and increased N2O emissions following
short-term (Moseman-Valtierra et al., 2011) and long-term exposure (Ji et al., 2015) to excess
nitrogen also suggests that nutrient enriched salt marshes may be a net source of N2O. Further,
results from agricultural soils demonstrate increase N2O emissions from fertilized fields,
61
however, the capacity of soils to be a sink or source of N2O is a balance between abiotic
conditions (i.e. pH) and biotic interactions (i.e. the presence of N2O scavengers; Jones et al.,
2014). Our results indicate that increased nutrient supply alters the carbon storage capacity of
salt marshes through enhanced decomposition through denitrification. If the stimulation of fungal
denitrification also increases N2O flux, this has important implications for the role of nutrient
enrichment in greenhouse gas emissions as well as carbon storage in coastal systems.
Acknowledgements: We would like to thank researchers of the TIDE Project (NSF
OCE0924287, OCE0923689, DEB0213767, DEB1354494, and OCE 1353140) for maintaining
the site and nutrient enrichment experiment. We would also like to thank Sarah Feinman and
members of the Bowen lab for their help in field collections. FTIR analysis was performed at the
Woods Hole Research Center in the lab of Jonathan Sanderman. Additional support was received
from the Plum Island LTER (NSF LTER OCE 0423565 and NSF OCE 1058747). This work was
funded by NSF award DEB 1350491 and DEB 1353140 to JLB and an NSF Research
Experience for Undergraduates Award DBI-1359241 to Dr. Rachel Skvirsky. All sequence data
from this study is available in the Sequence Read Archive under accession number SRP100756.
Literature Cited
Abarenkov, K., Henrik Nilsson R, Larsson KH, Alexander IJ, Eberhardt U, Erland S, et al.,
(2010) The UNITE database for molecular identification of fungi–recent updates and future
perspectives. New Phytol 186: 281-285.
Anderson MJ, (2001) A new method for non‐parametric multivariate analysis of variance. Aust
Ecol 26: 32-46.
Aronesty E, (2011) ea-utils: Command-line tools for processing biological sequencing data
http://code.google.com/p/ea-utils. Accessed September, 2016.
Bengtsson‐Palme J, Ryberg M, Hartmann M, Branco S, Wang Z, Godhe A, et al. (2013)
Improved software detection and extraction of ITS1 and ITS2 from ribosomal ITS sequences of
fungi and other eukaryotes for analysis of environmental sequencing data. Meth Ecol Evo 4: 914-
919.
62
Benner, R., Newell, S.Y., Maccubbin, A.E. and Hodson, R.E., 1984. Relative contributions of
bacteria and fungi to rates of degradation of lignocellulosic detritus in salt-marsh sediments.
Applied and Environmental Microbiology, 48(1), pp.36-40.
Blackwell M, (2011) The fungi: 1, 2, 3... 5.1 million species? Am J Bot 98 426-438.
Bokulich NA, Subramanian S, Faith JJ, Gevers D, Gordon JI, Knight R, et al. (2013) Quality-
filtering vastly improves diversity estimates from Illumina amplicon sequencing. Nat Meth 10:
57-59.
Borer ET, et al. (2014) Herbivores and nutrients control grassland plant diversity via light
limitation. Nature 508(7497):517–520.
Bowen JL, Crump BC, Deegan LA, Hobbie JE, (2009) Salt marsh sediment bacteria: their
distribution and response to external nutrient inputs. ISME J 3: 924-934.
Bowen JL, Ward BB, Morrison HG, Hobbie JE, Valiela I, Deegan LA, Sogin ML, (2011)
Microbial community composition in sediments resists perturbation by nutrient enrichment.
ISME J 5: 1540-1548.
Brin LD, Valiela I, Goehringer D, Howes B (2010) Nitrogen interception and export by
experimental salt marsh plots exposed to chronic nutrient addition. Mar Ecol Prog Ser 400: 3-17.
Burgaud G, Hue NT, Arzur D, Coton M, Perrier-Cornet JM, Jebbar M, Barbier G, (2015) Effects
of hydrostatic pressure on yeasts isolated from deepsea hydrothermal vents. Res Microbiol 166:
700e709.
Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. (2010)
QIIME allows analysis of high-throughput community sequencing data. Nat Meth 7: 335-336.
Chefetz B, Hader Y, Chen Y, (1998) Dissolved organic carbon fractions formed during
composting of municipal solid waste: Properties and significance. Acta Hydrochim Hydrobiol
26:172–179.
Craft C, Clough J, Ehman J, Joye S, Park R, Pennings S, et al., (2009) Forecasting the effects of
accelerated sea‐level rise on tidal marsh ecosystem services. Front Ecol Environ 7: 73-78.
Deegan LA, Bowen JL, Drake D, Fleeger JW, Friedrichs CT, Galvan KA, et al., (2007)
Susceptibility of salt marshes to nutrient enrichment and predator removal. Eco App 17: sp5.
Deegan LA, Johnson DS, Warren RS, Peterson BJ, Fleeger JW, Fagherazzi S, et al., (2012)
Coastal eutrophication as a driver of salt marsh loss. Nature 490: 388-392.
63
Dini-Andreote F, Pylro VS, Baldrian P, van Elsas JD, Salles JF, (2016) Ecological succession
reveals potential signatures of marine–terrestrial transition in salt marsh fungal communities.
ISME J 10: 1984-1997
Edgar RC, (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics
26: 2460-2461.
Ekschmitt, K., Kandeler, E., Poll, C., Brune, A., Buscot, F., Friedrich, M., Gleixner, G.,
Hartmann, A., Kästner, M., Marhan, S. and Miltner, A., 2008. Soil‐carbon preservation through
habitat constraints and biological limitations on decomposer activity. Journal of Plant Nutrition
and Soil Science, 171(1), pp.27-35.
Elahi R, O’Connor MI, Byrnes JE, Dunic J, Eriksson BK, Hensel MJ, Kearns PJ, (2015) Recent
trends in local-scale marine biodiversity reflect community structure and human impacts. Curr
Biol 25: 1938-1943.
Essington ME, (2004) Soil and water chemistry: An Integrative Approach. Taylor & Francis,
London, UK.
Fagherazzi S, Kirwan ML, Mudd SM, Guntenspergen GR, Temmerman S, D'Alpaos A, et al.,
(2012) Numerical models of salt marsh evolution: Ecological, geomorphic, and climatic factors.
Rev Geophys 50: RG1002.
Fox L, Valiela I, Kinney EL, (2012) Vegetation cover and elevation in long-term experimental
nutrient-enrichment plots in Great Sippewissett Salt Marsh, Cape Cod, Massachusetts:
implications for eutrophication and sea level rise. Estuar Coast 35: 445-458.
Galloway JN, Townsend AR, Erisman JW, Bekunda M, Cai Z, Freney JR, et al., (2008)
Transformation of the nitrogen cycle: recent trends, questions, and potential solutions. Science
320: 889-892.
Hammersley MR, Howes BL, (2005) Coupled nitrification-denitrification measured in situ in a
Spartina alterniflora marsh with a 15NH4+ tracer. Mar Ecol Prog Ser 299: 123-135.
Hirota M, Senga Y, Seike Y, Nohara S, Kunii H, (2007) Fluxes of carbon dioxide, methane and
nitrous oxide in two contrastive fringing zones of coastal lagoon, Lake Nakaumi, Japan.
Chemosphere 68: 597-603.
Hopkinson C, Giblin AE, (2008) Nitrogen dynamics in salt marsh ecosystems. Nitrogen in the
marine environment. Academic, 977-1022.
64
Howes BL, Weiskel PK, Geohringer DD, Teal JM, (1996) Interception of freshwater and
nitrogen transport from uplands to coastal waters: the role of saltmarshes. Estuarine shores:
evolution, environments and human alterations. John Wiley & Sons, 287-310.
Hsu JH, Lo SL, (1999) Chemical and spectroscopic analysis of organic matter transformations
during composting of pig manure. Environ Pollut 104: 189–196.
Ji Q, Babbin AR, Peng X, Bowen JL, Ward BB (2015) Nitrogen substrate–dependent nitrous
oxide cycling in salt marsh sediments. J Mar Res 73: 71-92.
Johnson DS, Warren RS, Deegan LA, Mozdzer TJ (2016) Saltmarsh plant responses to
eutrophication. Eco Apps 26: 2647-2659.
Kearns PJ, Angell JH, Feinman SG, Bowen JL (2015) Long-term nutrient addition differentially
alters community composition and diversity of genes that control nitrous oxide flux from salt
marsh sediments. Estuar Coast Shelf Sci 154: 39-47.
Kearns PJ, Angell JH, Howard EM, Deegan LA, Stanley RH, Bowen JL (2016) Nutrient
enrichment induces dormancy and decreases diversity of active bacteria in salt marsh sediments.
Nat Com 7: 12881.
Kenny C, Yamulki S, Blackwell M, Maltby E, French P, Birgand F, (2004) The release of
nitrous oxide from the intertidal zones of two European estuaries in response to increased
ammonium and nitrate loading. Water Air & Soil Pollut 4: 61-66.
Kirk, P.M., Cannon, P.F., Minter, D.W., Stalpers, J.A., 2008. Dictionary of the Fungi, tenth ed.
CABI, Wallingford, UK
Kirwan ML, Mudd SM, (2012) Response of salt-marsh carbon accumulation to climate change.
Nature 489: 550-553.
Kis-Papo T, (2005) Marine fungal communities. In: Dighton J, White JF, Oudemans P, Eds. The
Fungal Community: its Organization and Role in the Ecosystem. Taylor & Francis, Boca Raton,
FL, 61-92
Koop-Jakobsen K, Giblin AE (2010) The effect of increased nitrate loading on nitrate reduction
via denitrification and DNRA in salt marsh sediments. Limnol Oceano 55: 789-802.
Kozich JJ, Westcott SL, Baxter NT, Highlander SK, Schloss PD (2013) Development of a dual-
index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the
MiSeq Illumina sequencing platform. Appl Environ Microbiol 79: 5112-5120.
65
Kristensen E, Ahmed SI, Devol AH (1995) Aerobic and anaerobic decomposition of organic
matter in marine sediment: which is fastest? Limnol Oceano 40: 1430-1437.
Leff JW, Jones SE, Prober SM, Barberán A, Borer ET, Firn JL, et al., (2015) Consistent
responses of soil microbial communities to elevated nutrient inputs in grasslands across the
globe. Proc Natl Sci USA 112:10967-10972.
Levine JM, Brewer JS, Bertness MD, (1998) Nutrients, competition and plant zonation in a New
England salt marsh. J Ecol 86: 285-292.
Lin X, Feng Y, Zhang H, Chen R, Wang J, Zhang J, Chu H, (2012) Long-term balanced
fertilization decreases arbuscular mycorrhizal fungal diversity in an arable soil in North China
revealed by 454 pyrosequencing. Environ Sci Tech 46: 5764-5771.
Maeda K, Spor A, Edel-Hermann V, Heraud C, Breuil MC, Bizouard F, et al., (2015) N2O
production, a widespread trait in fungi. Sci Rep 5: 9697.
Mcleod E, Chmura GL, Bouillon S, Salm R, Björk M, Duarte CM, et al., (2011) A blueprint for
blue carbon: toward an improved understanding of the role of vegetated coastal habitats in
sequestering CO2. Front Ecol Environ 9: 552-560.
Middelburg JJ, Klaver G, Nieuwenhuize J, Markusse RM, Vlug T, Nat FJ (1995) Nitrous oxide
emissions from estuarine intertidal sediments. Hydrobiologia 311: 43-55.
Moseman-Valtierra S, Gonzalez R, Kroeger KD, Tang J, Chao WC, Crusius J, et al., (2011)
Short-term nitrogen additions can shift a coastal wetland from a sink to a source of N2O. Atmos
Envir 45: 4390-4397.
Nagano Y, Nagahama T, Hatada Y, Nunoura T, Takami H, Miyazaki J, Takai K,Horikoshi K,
(2010) Fungal diversity in deep-sea sediments e the presence of novel fungal groups. Fungal
Ecol 3: 316e325.
Nixon SW, Oviatt CA, Frithsen J, Sullivan B (1986) Nutrients and the productivity of estuarine
and coastal marine ecosystems. Afr J Aquat Sci 12: 43-71.
O'Brien HE, Parrent JL, Jackson JA, Moncalvo JM, Vilgalys R, (2005) Fungal community
analysis by large-scale sequencing of environmental samples. Appl Environ Microbiol 71: 5544-
5550.
Oksanen J, Kindt R, Legendre P, O’Hara B, Stevens MHH, Oksanen MJ, et al., (2007) The
vegan package. Community ecology package, 10.
66
Paerl HW, (1997) Coastal eutrophication and harmful algal blooms: Importance of atmospheric
deposition and groundwater as “new” nitrogen and other nutrient sources. Limnol Oceano 42:
1154-1165.
Parikh SJ, Goyne KW, Margenot AJ, Mukome FND, Calderón FJ, (2014) Soil chemical insights
provided through vibrational spectroscopy. In: D.L. Sparks, editor, Advances in Agronomy.
Academic Press, Waltham, MA. p. 1-148.
Peng X, Ji Q, Angell JH, Kearns PJ, Yang HJ, Bowen JL, Ward BB, (2016) Long‐term
fertilization alters the relative importance of nitrate reduction pathways in salt marsh sediments.
J Geophys Res Biogeoscience 121: 2082-2095.
Pennings SC, Grant MB, Bertness MD, (2005) Plant zonation in low‐latitude salt marshes:
disentangling the roles of flooding, salinity and competition. J Ecol 93: 159-167.
Picard KT (2017) Coastal marine habitats harbor novel early-diverging fungal diversity. Fungal
Ecol 25: 1-13.
Rabalais NN, Turner RE, Wiseman Jr WJ, (2002) Gulf of Mexico hypoxia, aka" The dead zone".
Annu Rev Ecol Evol Syst, 235-263.
Ravishankara AR, Daniel JS, Portmann RW, (2009) Nitrous oxide (N2O): the dominant ozone-
depleting substance emitted in the 21st century. Science 326: 123-125.
Reed DJ (1990) The impact of sea-level rise on coastal salt marshes. Prog Phys Geog 14: 465-
481.
Richards TA, Jones MD, Leonard G, Bass D, (2012) Marine fungi: their ecology and molecular
diversity. Ann Rev Mar Sci 4: 495-522.
Solomon, D., Lehmann, J., Kinyangi, J., Amelung, W., Lobe, I., Pell, A., Riha, S., Ngoze, S.,
Verchot, L.O.U., Mbugua, D. and Skjemstad, J.A.N., 2007. Long‐term impacts of anthropogenic
perturbations on dynamics and speciation of organic carbon in tropical forest and subtropical
grassland ecosystems. Global Change Biology, 13(2), pp.511-530.
Stoeck T, Hayward B, Taylor GT, Varela R, Epstein SS, (2006) A multiple PCR primer
approach to access the microeukaryotic diversity in environmental samples. Protist 157: 31e43.
Taylor JD, Cunliffe M, (2016) Multi-year assessment of coastal planktonic fungi reveals
environmental drivers of diversity and abundance. ISME J 10: 2118-2128.
67
Torta, L., Lo Piccolo, S., Piazza, G., Burruano, S., Colombo, P., Ottonello, D., Perrone, R., Di
Maida, G., Pirrotta, M., Tomasello, A. and Calvo, S., 2015. Lulwoana sp., a dark septate
endophyte in roots of Posidonia oceanica (L.) Delile seagrass. Plant Biology, 17(2), pp.505-511.
Valiela I, Teal JM, Sass W, (1973) Nutrient retention in salt marsh plots experimentally fertilized
with sewage sludge. Estuar and Coast Mar Sci 1: 261-269.
Valiela I, Teal JM, (1979) The nitrogen budget of a salt marsh ecosystem. Nature 280: 652-656.
Valiela I, Teal JM, Persson NY (1976) Production and dynamics of experimentally enriched salt
marsh vegetation: belowground biomass. Limnol Oceano 21: 245-252.
Valiela I, Foreman K, LaMontagne M, Hersh D, Costa J, Peckol P, et al., (1992) Couplings of
watersheds and coastal waters: sources and consequences of nutrient enrichment in Waquoit
Bay, Massachusetts. Estuaries 15: 443-457.
Valiela I, Teal JM, Allen SD, Van Etten R, Goehringer D, Volkmann S, (1985) Decomposition
in salt marsh ecosystems: the phases and major factors affecting disappearance of above-ground
organic matter. J Exp Mar Biol Ecol 89: 29-54.
Veum K, Goyne K, Kremer R, Miles R, Sudduth K (2014) Biological indicators of soil quality
and soil organic matter characteristics in an agricultural management continuum.
Biogeochemistry 117: 81–99.
Walters W, Hyde ER, Berg-Lyons D, Ackermann G, Humphrey G, Parada A et al., (2016)
Improved Bacterial 16S rRNA Gene (V4 and V4-5) and Fungal Internal Transcribed Spacer
Marker Gene Primers for Microbial Community Surveys. mSystems 1: e00009-15.
Zheng Y, Chen L, Luo CY, Zhang ZH, Wang SP, Guo LD, (2016) Plant identity exerts stronger
effect than fertilization on soil arbuscular mycorrhizal fungi in a sown pasture. Micro Ecol 72:
647-658.
Zhou J, Jiang X, Zhou B, Zhao B, Ma M, Guan D et al., (2016) Thirty-four years of nitrogen
fertilization decreases fungal diversity and alters fungal community composition in black soil in
northeast China. Soil Biol Biochem 95: 135-143.
Zumft WG (1997). Cell biology and molecular basis of denitrification. Microbiol Mol Biol Rev
61: 533-616.
68
Tables and figures
Table 1- Wavelengths and functional assignments used in this study based on Parikh et al.,
(2014). δ = bending vibration, ʋ= stretching vibration, ʋas =asymmetric stretching vibration, and
ʋs = symmetric stretching vibration.
Wavelength, cm-1 Assignment
3402 O-H and v N-H, hydroxyl amine
2924 aliphatic ʋas(C-H)
2850 aliphatic ʋs(C-H)
1650 aromatic ʋ(C = C)
1470 aliphatic δ(C-H)
1405 aliphatic δ(C-H)
920 aromatic δ(C-H)
840 aromatic δ(C-H), less substituted
69
Figure 1: Principal coordinates analysis (A) based on Bray-Curtis similarity values for fertilized
(red) and reference marshes (blue). Boxplot of paired Bray-Curtis similarity values (B) for all
four habitats comparing similarity between nutrient enriched and reference samples. Boxes
represent 25-75% quartiles and the solid black line indicates the median value. Letters in (B)
indicate significantly different categories (p<0.05) assessed by a Tukey HSD test.
70
Figure 2: Box and whisker plots of fungal Shannon Diversity (A) and log10 ITS abundance (B)
assessed by quantitative PCR for fertilized (red) and reference (blue) habitats. Boxes represent
25-75% quartiles and the solid black line indicates the median value. Letters indicate
significantly different (p<0.05) categories as assessed by a Tukey HSD test.
71
Figure 3: Stacked bar plot of the average relative abundance of the top 20 fungal families in the
four marsh habitats across reference and fertilized creeks. Data shown comprises 90% of all
sequences and the category ‘other’ contains all remaining taxa. k= kingdom, c=class, o=order,
f=family.
72
Figure 4: Log10 abundance of significantly different OTUs (Kruskal-Wallis test, Benjamini-
Hochberg corrected p< 0.05, abundance >50) between fertilized and reference sediments. OTUs
are identified to the highest possible taxonomic resolution. k=kingdom, p=phylum, o=order, and
f=family.
73
Figure 5: Box and whisker plot of the percentage of fungal denitrifiers present in fertilized (red)
and reference (blue) habitats. Fungal denitrifiers were recently identified by Maeda et al., (2015).
Boxes represent 25-75% quartiles and the solid black line indicates the median value. Letters
indicate significantly different (p<0.05) categories as assessed by a Tukey HSD test.
74
Figure 6: Biplot of the principal coordinates axis 1 values from Fig. 1 versus percent carbon for
reference (blue) and fertilized (red) marshes (A). Plots of Shannon Diversity (B) and log10 ITS
copy number (C) for reference marshes. Colored lines in A-C are ANCOVA regressions and the
equations and statistics for each line are denoted in the panel.
75
Figure 7: Correspondence analysis (A) of FTIR spectra for fertilized (red) and reference (blue)
habitats. Box and whisker plots of Index I which is a metric of decomposition (B) and the ratio
of hydroxyl amines to three wavelengths of aromatic carbon (C) which is a metric of carbon
lability. Boxes represent 25-75% quartiles and the solid black line indicates the median value.
Letters in (B) and (C) indicate significantly different (p<0.05) categories as assessed by a Tukey
HSD test.
76
Chapter 2: Supplemental Information
SI Figure 1: Schematic of a typical Plum Island marsh depicting the 4 habitats sampled in this
study. Modified from Johnson et al., (2016).
77
SI Figure 2: Plot of average Fourier Transform Spectroscopy spectra of fertilized (red) and
reference (blue) for mudflat (A), tall Spartina alterniflora (B), S. patens (C), and short S.
alterniflora (D) habitats.
78
Chapter 3: The effect of short-term, diel changes in environmental conditions on active
microbial communities in a salt marsh pond
Abstract
Microbial communities play key roles in biogeochemical cycles across the planet and their
composition and the factors that structure them are well documented. However, how changes in
abiotic conditions affect the active proportion of the microbial community that is responsible for
the delivery of key ecosystem services is poorly understood. Salt marshes, in particular salt
marsh ponds, are highly dynamic habitats with abiotic conditions in the pond water that fluctuate
on daily cycles. To determine how diurnally driven changes in abiotic conditions affect active
microbial communities we sampled a single salt marsh pond every four hours over two diel
cycles, sampling two dynamically different habitats, the pond sediment and overlying water. We
assessed abiotic conditions and the total and active microbial communities using high-throughput
sequencing of the 16S rRNA gene and 16S rRNA. Sediments displayed no discernable pattern
and low variation in abiotic conditions, leading to stable active microbial communities.
However, the cyclical, rapidly changing abiotic conditions in the overlying water resulted in
large swings in the active microbial community and in the percent of inactive taxa. Our data
suggest that changes in environmental condition over short time periods alter the structure of
active microbial communities. Further, our data show that the most abundant active taxa in the
overlying water were rare (<1% total abundance) suggesting that under environmental change,
rare taxa can disproportionally contribute to the activity of microbial communities.
Introduction
Microbial communities face considerable physiological challenges from changing abiotic
conditions and the way microbes respond to these changes can have profound effects on
ecosystem function. Broadscale spatial and temporal drivers often control microbial composition
(Lozupone and Knight 2007, Gilbert et al. 2012), however, variation in local environmental
conditions can confer changes in microbial community composition, activity, and geochemical
output (Alison and Martiny 2008, Shade et al. 2012, Griffiths and Philippot 2013, Reed and
79
Martiny 2013). Under some circumstances, however, microbial community composition may
remain resistant to perturbations (Bowen et al. 2011, Kearns et al. 2016), despite differences in
microbial activity or geochemistry (Bowen et al. 2011, Kearns et al. 2016, Koop-Jakobson and
Giblin 2010). While variation in community composition can lead to changes in function (Reed
and Martiny 2013), the correlation between composition and function can also be disconnected
(Frossard et al. 2013, Bowen et al. 2014, Balser and Firestone 2005, Strickland et al. 2009) due
to offsets in timing between gene expression and associated function, or because the end product
of microbial activity can be consumed prior to being detected (i.e. cryptic cycling). Functional
redundancy, where multiple organisms perform the same function, can also result in a change in
community structure with no net change in function (Alison and Martiny 2008). Finally, inactive
or dormant microbes may mask our ability to determine microbial responses to a changing
environmental conditions (e.g. Bowen et al. 2011, Kearns et al. 2016).
Dormancy is a strategy some microorganisms use to persist in their environment during
unfavorable conditions (Lennon and Jones 2011). Dormant microbes exist in a wide range of
ecosystems, however their proportion of the total bacterial community differs (Lennon and Jones
2011) and can limit our ability to detect community response to environmental change.
Ecosystems that have high proportions of dormant taxa, such as soils and sediments, possess a
large degree of spatial and temporal heterogeneity (Lennon and Jones 2011, Kearns et al. 2016).
Additional controls on the proportion of inactive or dormant cells include changes in abiotic
conditions such as salinity (Aanderud et al. 2016), and excess nutrients (Kearns et al. 2016),
however nutrient limitation also appears to increase dormancy in some systems (Jones and
Lennon 2010, Aanderud et al. 2016). Ultimately, the controls over dormancy and the role of
environmental change on the proportion of dormant cells remains unclear, resulting in a pressing
need to understand the role of the environment in maintenance and resuscitation of dormant taxa.
Salt marshes provide a dynamic natural experiment to test hypotheses regarding how
varying environmental conditions influence the activity of micro-organisms. Salt marshes are
highly productive, tidally flooded coastal grassland ecosystems that provide numerous services
to both the land and sea (Teal 1962, McLeod et al. 2011, Deegan 1993). Due to their low
elevation relative to mean tidal height (Nicholls et al. 1999), salt marshes are highly susceptible
to changes in sea level. Increases in sea levels can cause net shifts in marsh plant communities as
80
well as increased numbers of ponds, unvegetated areas that have lower elevation than the
surrounding marsh platform (Redfield 1972, Reed 2002, Day et al. 2011). Salt marsh ponds can
exchange water with the coastal ocean on very high flooding tides, however, ponds often
experience extended periods of disconnect from tidal flooding. During their isolation, the water
in marsh ponds can experience oscillations of oxygen and other constituents over short periods of
time (Smith and Able 2003, Day et al. 2011). Conversely, the sediments underlying ponds
remain relatively unperturbed on those time steps. Salt marsh ponds are very shallow (often <20
cm water depth), so sediments and overlying water represent highly connected, yet distinct
systems that are ideal for studying how environmental change alters microbial activity over short
time scales.
The activity of microbial cells has often been assessed using transcripts of the small
subunit of prokaryotic ribosomes (16S rRNA) due to its correlation with protein synthesis
(Kerkhof and Kemp 1999, Deuscher et al. 2006) and it has also been used to assess dormancy in
environmental samples (Jones and Lennon 2010). Despite the caveats to this approach
(Blazewicz et al. 2013), 16S rRNA can be useful as a relative measure of active microbial
community structure and how it varies over time or as a result of experimental manipulation
(Jones and Lennon 2010, Campbell et al. 2011, Aanderud et al. 2016). To assess the effect of
cyclical diurnal changes in environmental conditions on active microbial community
composition we sampled the sediment and overlying water in a salt marsh pond every four hours
over two diel cycles (48 hours). The abiotic conditions within the water change both cyclically
and rapidly and the sediment abiotic conditions, while variable, did not systematically change
allowing us to test the hypothesis that that cyclically dynamic abiotic conditions in the overlying
water (temperature, oxygen, salinity, pH, reactive nitrogen species) caused by high daytime
primary productivity, would drive strong changes in active community composition, while static
conditions in the sediment would promote stable active communities. We sequenced the 16S
rRNA gene and 16S rRNA to determine the effect of diel cycles on total and active community
composition respectively. Results indicate that there were distinct oscillations in abiotic
conditions in the overlying water column but no such oscillating pattern in the sediment pore
water abiotic conditions, despite variability in sediment abiotic conditions. Active water column
microbial communities displayed wide shifts in abundance and changes in community structure.
Conversely, the sediment communities, where abiotic conditions remained relatively constant,
81
displayed no shift in the abundance or community structure of active taxa in response to diel
cycles. These results suggest that diel cycles structure the active portion of water column
communities and that dynamic conditions can promote rapid changes in the structure of active
microbial communities.
Methods
Study Site and Sample Collection
We sampled a salt marsh pond located at the Plum Island Ecosystems Long Term
Ecological Research Site in Rowley, MA, USA (42.759 N, 70.891 W). The pond (located at
42.741 N, -70.831 W) was ~20 cm in depth and approximately 7120 m2 in area. During the
course of the experiment and two days prior to our sampling, the pond was isolated from tidal
flooding due to low tidal amplitude. Sampling began at 10 AM on June 30, 2014 and ended at 10
AM on July 2, 2014, capturing two full diel cycles. On the days sampled, the sun rose between
5:09 and 5:15am and set between 8:25 and 8:30pm. The weather was clear with nominal wind
speeds (<1 m s-1) and the air temperature ranged from a low of 17°C to a high of 31°C (mean=
24°C)..To account for spatial variability within the pond, we sampled three different locations
that were approximately 15 meters apart and at each site, we sampled within 1 meter of a central
point.
Every two hours over the course of two days, we collected environmental parameters
(dissolved oxygen, salinity, pH, temperature) with a Hanna HI9828 multiparameter meter
(Geoscientific ltd, Vancouver, British Columbia, Canada) and we collected and filtered 50-mL of
water from each location for nutrient analysis. The water was passed through a 0.46 µm
Whatman® GF/F filter into sterile 50-mL centrifuge tubes and stored on dry ice. Additionally,
every four hours one liter of water was filtered onto a 0.22 µm Millipore SterivexTM filter from
each site and flash frozen in liquid nitrogen, leading to a total of 39 water samples. We collected
39 surface sediment (top 2-cm) samples from all three locations within the pond using a 15-cm
inner-diameter PVC corer. Sediments from each site were individually homogenized in a sterile
centrifuge tube and subsamples were placed into cryovials for storage on liquid nitrogen. The
remaining sediment was stored on dry ice for determination of carbon and nitrogen content.
Nutrient Analyses
82
Nitrate, ammonium, and phosphate concentrations in sediment pore water and the
overlying water column were analyzed on a Lachat Flow 118 Injection Analyzer (Hach,
Loveland, CO, USA) based on QuikChem methods 31-107-04-1-E 119, 31-107-06-1-B, E31-
115-01-1-W, and E31-114-27-1-E respectively at the UMass Boston Environmental Analytical
Facility. The lower limit of detection for the methods was 0.60 µM, 1.45 µM, 0.25, and 2.5 µM
respectively. Due to the low volume of sediment pore water, nitrate concentrations were
measured by chemiluminescence after vanadium reduction (Garside 1982, Braman and Hendrix
1989) on a Teledyne NOx analyzer (Thousand Oaks, CA, USA). Elemental composition (percent
nitrogen and carbon) was measured on sediments dried at 50°C for three days, using a Perkin
Elmer 2400 Series Elemental Analyzer (Waltham, MA, USA) following standard procedures. To
understand the variability of abiotic conditions within the pond water and porewater, we
calculated the coefficient of variation.
Nucleic Acid Extraction and Reverse Transcription
Nucleic acids from sediments were extracted with the MoBio PowerSoil Total RNA
Isolation Kit (Carlsbad, CA, USA) from ~1.5 g of sediment following manufacturer’s
instructions. Nucleic acids were extracted from water filters using the MoBio PowerWater Kit
following manufacturer’s instructions. RNA was checked for DNA contamination with general
bacterial primers and any DNA contamination was removed using DNAse I (New England
BioLabs, Ipswich, MA, USA) following the manufacturer’s instructions. RNA was reverse
transcribed to cDNA using the Invitrogen Superscript RT III cDNA synthesis kit (Carlsbad, CA,
USA) following manufacturer’s instructions for random hexamers. cDNA synthesis was
confirmed with PCR using general bacterial primers and verified on a 1.5% agarose gel stained
with ethidium bromide.
PCR and Sequencing
DNA and RNA was quantified with Picogreen and Ribogreen (Invitrogen) kits
respectively and nucleic acids were normalized to 1 ng µl-1 for all PCR reactions. Samples were
prepared for sequencing on an Illumina MiSeq following the protocol outlined in Caporaso et al.
(2011). We used general bacterial primers 515F and 806R (Caporaso et al. 2011), with Illumina
adaptors and individual 12-bp GoLay barcodes attached to reverse primers, to amplify a section
of the V4 hypervariable region of the 16S rRNA gene. PCR reactions were performed in
triplicate, verified with gel electrophoresis, and fragments were excised and purified using the
83
Qiagen QIAquick gel extraction kit (Qiagen, Valencia, CA, USA). Samples were quantified
fluorometrically on a Qubit (ThermoFisher, Waltham, MA, USA) and pooled in equal molar
concentrations for sequencing on the Illumina MiSeq (San Diego, CA, USA) platform for paired-
end 151 bp sequencing. All sequencing was performed at the University of Massachusetts
Boston using V2 chemistry.
Quantitative PCR
Quantative PCR (qPCR) was performed to assess 16S rRNA copy number for both the
active and total microbial communities. All DNA, cDNA, and standards prepared from purified
PCR product we quantified with a Qubit (ThermoFisher). DNA from each sample was
normalized to 3 ng µl-1 and serial dilutions of standards were prepared. DNA and cDNA were
amplified in triplicate reactions, along with standards and internal controls on a Strategene MX-
3005p quantitative thermocycler (Stratagene, La Jolla, CA, USA) with the primer pair 357F and
519R (Biddle et al. 2008) using conditions described previously (Bowen et al. 2011). All qPCR
reactions were verified for proper product formation with gel electrophoresis and melt curves.
All standard curves had R2 values >0.99 and qPCR efficiency ranged from 90-103%. To test
significant differences in abundance of 16S rRNA we used an ANOVA in R (R Core Team
2011).
Sequence and Data Analysis
Paired end reads were joined with fastq-join (Aronsety 2011) using default parameters
and sequences were demultiplexed and quality filtered in QIIME (version 1.9; Caporaso et al.
2010) following the guidelines recommended by Bokulich et al. (2013). Sequences were checked
for chimeras using USEARCH in de novo mode and subsequently removed (Edgar 2011).
Following quality filtering, we retained a total of 3.87 million sequences. Operational taxonomic
units (OTUs) were clustered at 97% sequence identity with Swarm (Mahé et al. 2013) in QIIME
and OTUs appearing only once across the dataset were discarded. Taxonomy was assigned using
uclust (Edgar et al. 2011) against the GreenGenes database (version 13.5). All sequences
matching archaea and chloroplasts were filtered from the dataset. Sequences were aligned with
PyNast (Caporaso et al. 2010) and a phylogenetic tree was constructed with fasttree (Price et al.
2010).
84
We calculated beta diversity on normalized OTU tables using weighted UniFrac
(Lozupone and Knight 2005) in QIIME. Significant differences in groups were assessed with a
non-parametric MANOVA (Anderson 2002) with 10,000 permutations. Alpha diversity was
calculated on a rarefied OTU table (10,000 restarts, steps of 50) normalized to the lowest
sampling depth (11,596 sequences). To assess heterogeneity in the sediment communities and
determine if overdispersion was impeding our ability to detect changes in sediment communities
we performed a regression-based overdispersion test (Cameron and Trivedi 1990) in R. We
calculated Shannon Diversity and phylogenetic diversity and assessed significant differences in
diversity with a one-way ANOVA in R (R Core Team 2011). To assess the degree of inactivity
among taxa in our samples we calculated the 16S rRNA:16S rRNA gene ratio for each OTU in
each sample and defined a taxon as active with a ratio >1 (Lennon and Jones 2010). To assess
the contribution of the rare biosphere to the active communities of the sediment and water we
used a Chi-squared test to determine the differences in abundance between OTUs in the total
(expected) and active (observed) microbial communities using taxa with average abundance in
the total microbial community of <1%. Further, we calculated non-parametric Spearman
correlations between 16S rRNA gene and 16S rRNA profiles to examine how the abundance of
taxa in the total communities correlates to their abundance in the active community.
Results
Environmental Conditions
Abiotic conditions in the marsh pond sediment porewater (Fig. 1) did not vary
appreciably over the time course of our experiment. The coefficient of variation (the ratio of the
standard deviation to the mean) was typically low for environmental parameters (Fig. 1), ranging
from 0.01 to 0.03, suggesting low variability in environmental conditions in the sediment over
the 48 hour experiment. While there was slight variation in abiotic conditions, this variability
was not due to systematic temporal effects. Abiotic parameters in the overlying water column,
however, did vary with the time of day (Fig. 1). Temperature, pH, and oxygen had high levels
during the day and decreased levels at night. Ammonium and nitrate displayed the inverse
pattern, with high concentrations at night and low or undetectable concentrations during the day.
Water column salinity, while variable, only changed by ~0.5 ppt and phosphate was stable over
time. The coefficient of variation of environmental conditions in the water (Fig. 1) ranged from
85
0.9 to 1.97 and was significantly higher than the variation in the sediments (t-test, T=14.90,
p<0.002), providing evidence that environmental conditions in the water were relatively more
temporally dynamic than in the sediment.
Community Composition and Abundance
The total (16S rRNA gene) and active (16S rRNA) communities in the pond sediment
were significantly different from each other (PERMANOVA, F(1,78)=12.97, p<0.001; Fig. 2A),
but neither the total (F(4,35)=4.45, p=0.60) nor potentially active (F(4,35)=6.04, p=0.56) sediment
communities displayed significant short-term (F(5,34)=2.43, p=0.91) temporal or spatial
(F(2,37)=0.90, p=0.64) shifts in composition and the variation that exists is nominal for total (mean
weighted UniFrac similarity=96.04%) and active communities (mean= 94.56%). Further, a
overdispersion test indicated no significant (p>0.56) effects of time or site on sediment
composition. As with sediment communities, water column active and total microbial
communities were significantly different from each other (F(1,78)=19.33, p<0.001; Fig. 2B). The
total community in the overlying water also remained stable over the 48-hour period, displaying
no significant spatial (F(3,36)=9.45, p=0.90) or temporal effects (F(4,34)=11.43, p=0.80), and with
an average weighted UniFrac similarity of 97.5%. The potentially active community, however,
displayed consistent and recurring diurnal changes over the 48 hours of our experiment,
displaying five significantly different clusters (F(4,34)=19.21, p<0.001) corresponding to time of
day with no evidence of spatial effects with average similarity of 52.43%. A plot of the ratio of
the abundance of 16S rRNA:16S rRNA gene, as assessed by qPCR (SI Fig. 1) revealed a
significantly higher ratio in sediment than water communities (ANOVA; F=29.45, p<0.001) but
neither ratio varied significantly over the course of the experiment in the sediment or the water.
Diversity and Activity
Shannon Diversity in the active and total communities remained constant in the sediment
over the 48-hour period (SI Fig. 2). Both the sediment and water column total community had
higher Shannon Diversity than the potentially active community. The water column active
community, however, had significantly different Shannon Diversity values at different times of
day. Patterns in the percent of inactive taxa (16S rRNA:16S rRNA gene <1) varied between
sediment and water (Fig. 3). The percent of inactive taxa in the sediment ranged from 45 to 55%
of the community (Fig. 3A) and did not vary significantly during the duration of this experiment
or between sampling locations. Inactivity in the water column, by contrast, ranged from 50 to
86
65% of the community, and was temporally variable (ANOVA; p<0.001, F(4,24)=12.43). The
percentage of inactive taxa was significantly higher in the water than the sediment (ANOVA;
p<0.01, F(1,76)=12.43).
Taxonomic Composition
The taxonomic composition of the total sediment community did not vary over the 48-hr
experiment (SI Fig. 3). This community was comprised of largely anaerobic or facultatively
anaerobic orders including Chromatiales (~17%), Desulfobacterales (7%), Alteromonadales
(12%), Myxococcales (9%), Bacteroidales (8%), and Cytophagales (7.5%). The potentially
active sediment community (Fig. 4A) also did not vary in its structure over time and was largely
composed of several largely anaerobic orders such as Chromatiales (~19%), Alteromonadales
(~7%), Desulfobacterales (~9%), and Myxococcales (8%). Phototrophic taxa such as
Cyanobacteria (~15%) were present in the potentially active sediment microbial community
throughout the course of the experiment.
The taxonomic composition of the total water community (SI Fig. 4) was consistent
through time and was largely composed of the orders Fusobacterales (~12%), Actinomycetales
(~7%), Thiotrichales (10%), Myxococcales (9%), and Altermonadales (11%). A stacked bar plot
of the active communities of the water column displayed dynamic shifts in taxonomic
composition (Fig. 4B). Peak photosynthetic hours (10 AM/2 PM) were dominated by an
unclassified Cyanobacterial order (~60%) in addition to the Alphaproteobacterial order
Rickettsiales (~13%). Four hours later at 6 PM, the water column active microbial community
was dominated by the Gammaproteobacterial orders Virbionales (~70%) and Pseudomonadales
(~10%). At 10 PM, the first time point after sunset, the composition of the active community was
primarily comprised of the orders Vibrionales (~15%), Flavobacterales (~10%), Pseudomondales
(~7%), and Actinomycetales (~22%). When oxygen was at its lowest at 2 AM, the active water
column community was dominated by the order Bacillales (~50%), with a smaller contribution
from the order Vibrionales (~15%). Finally, at 6 AM, the time point that harbored the largest
diversity of active taxa, the community was comprised of the orders Flavobacterales (~13%),
Unclassified Cyanobacteria (~7%), and Burkholderiales (~17%).
Activity of Dominant Taxonomic Groups
We further explored patterns among active taxa by calculating the 16S rRNA:16S rRNA
gene ratio for the most active taxa (16S ratio >15) in the overlying water (Fig. 5 And aggregating
87
the data at the order level. A heatmap of this ratio revealed numerous taxa with activity that was
restricted to few times of the day (unclassified Cyanobacteria, Rickettsiales, Acidimicrobiales,
Bacillales) suggesting their activity was limited by environmental conditions or competition with
other taxa. Other taxa were active more frequently (Vibrionales, Burkholderiales,
Actinomycetales, Marinicellales) suggesting less stringent growth requirements. Further, certain
taxa appear to be active/inactive in concert with one another (unclassified Cyanobacteria and
Rickettsiales) possibly suggesting a symbiotic relationship, niche overlap, or a similar response
to environmental queues. A heat map of the most active sediment taxa (SI Fig. 5) generally
displayed a consistent ratio for the most active taxa over the course of the 48-hr experiment.
Rank Abundance and the Rare Biosphere
To assess the role that abundance of taxa in the total community plays in their
concomitant abundance in sediment and water active communities we created bi-plots comparing
the abundance of OTUs in the 16S rRNA to their abundance in the 16S rRNA gene (Fig. 6). The
abundance of sediment taxa in the 16S rRNA was strongly correlated (Spearman’s ρ=0.85) to
their abundance in the 16S rRNA gene (Fig. 7A), as highlighted by the highly abundant OTUs
from the order Chromatiales (0.15-0.20 relative abundance), which was similar in abundance in
the total and active communities. Water communities, however, displayed a disconnect in the
abundance in the 16S rRNA and 16S rRNA gene (Spearman’s ρ =0.31) and that disconnect was
largely due to several taxa, including those from the order Vibrionales, Bacillales, Rickettsiales,
and unclassified Cyanobacteria, that were in low abundance in the 16S rRNA gene compared to
the 16S rRNA, suggesting a role for rare taxa. We next explored the role of the rare biosphere
by constructing plots comparing 16S rRNA gene rank of OTUs versus 16S rRNA abundance
rank-abundance curves (SI Fig. 7). Rank-abundance plots allows us to both explore the
correlation in abundance in the total and active communities and to determine the contribution of
rare taxa to active community composition. The sediment community rank abundance curve (SI
Fig. 7A) indicated that the most abundant taxa in the total community were also the most active
taxa. Indeed, no taxa in the sediment rare biosphere (<1% total abundance) significantly varied in
abundance between the total and active microbial communities (Chi-squared test, p>0.5).
Communities in the overlying water, however, had vastly different rank-abundance curves
relative to the sediment (Fig. 7B). Typically, the most abundant taxa in the active water column
community were in very low abundance in the total community (<1% total abundance) and of
88
the rare taxa, 15% displayed significantly higher abundance in the active community relative to
the total community (Chi-square test, p<0.05) suggesting a greater contribution of the rare
biosphere to the structure of the active water community than the sediment.
Discussion
Our study examined the effect of short-term changes in environmental conditions on the
total and active microbial communities in a salt marsh pond. Our analysis revealed differential
responses of sediment and water column microbial communities to diel changes in
environmental conditions. Sediments, while having significantly different total and active
communities (Fig. 2A), displayed no temporal changes in active microbial community
composition or diversity. The lack of change in the active sediment communities may be due to
the short duration of this study (48 hrs), which may be out of step with the duration of time it
takes abiotic conditions in the sediments to significantly vary. In addition, stochastic assembly
processes (Stegen et al. 2012, Nemergut et al. 2013) may mask our ability to detect changes in
sediment communities. However, the striking stability of the taxonomic composition of the total
and active sediment communities (Fig. 4, SI Fig. 3) and high degree of similarity between
samples suggests this is not the case. Coastal sediments display sharp, shallow gradients in
numerous abiotic condition (Canfield et al. 1993) that can confer gradual changes to microbial
community composition (Armitage et al. 2012). Our sampling method (top 2-cm) integrates over
these gradients and may mask potential fine-scale changes that occur over shallow spatial scales.
However, the sampling technique employed here (top 2 cm of surface sediment) was previously
(Kearns et al. 2016) used to elucidate the dramatic effect of nitrogen enrichment on active salt
marsh sediment bacterial communities, suggesting that these methods should be robust to detect
integrated large-scale changes in the active bacterial community within the pond. Finally, we’ve
documented highly abundant taxa from the Chromatiales in the sediment (~20% relative
abundance) and water (~1.5% relative abundance) suggesting that as in vegetated marsh areas
(Bolhuis and Stal 2011, Kearns et al. 2016), anoxygenic photosynthesizers from the order
Chromatiales play an important role in the ecology of salt marsh ponds.
Conversely, the water column displayed a highly variable active microbial community
(Fig. 2B) and a diel change in active diversity (SI Fig. 2B). Our results suggest that cyclically
changing environmental conditions can confer rapid shifts in the active microbial community,
89
while the relatively unchanging conditions found in the pond sediments promote more consistent
structure across time for both active and total bacterial communities. Changes in active
community composition can be correlated with changes in transcriptional activity (Gilbert et al.
2010) suggesting that relatively static habitats, like pond sediments, likely promote consistent
metabolic processes while the microbial communities in dynamic habitats likely displayed
variable metabolic output due to environmental changes. In the water column, we observed
There were large variations in the percentage of inactive taxa (Fig. 3B; 16S ratio <1) in
the overlying water. Of the 352 microbial orders present within the active community of the
overlying water, 62% (n=217) never achieve a 16S rRNA to 16S rRNA gene ratio (hereafter,
16S ratio) greater than one, suggesting these taxa were likely dormant within this system. Several
orders (n=52) had 16S ratios that oscillated between <1 and >10 suggesting an increase and
decrease in their activity over time in response to environmental conditions. These taxa were
likely not in a true state of dormancy during times when their ratio was less than one since the
cellular investment dormancy takes (Lennon and Jones 2011) is a longer-term ecological
strategy. A final group of taxa (n=82), had 16S ratios that cycle between less than one and five
and their status as inactive (short durations of time with ratios less than one) or dormant (longer
durations of time with ratios less than one) remains unclear because several taxa have been
shown to need a 16S ratio greater than one to divide (Blazewicz et al. 2013). However, cell
division is not the only form of activity a microbe can take, instead microbes may be performing
biogeochemical processes or other cellular processes such as repair rather than dividing
(Kirchman 2016). Our results highlight the difficulty in interpreting 16S ratios from
metagenomic data, in particular for taxa with ratios that change rapidly. Further work on
microbial dormancy is needed to better understand what the 16S ratio means for phylogenetically
widespread taxa and to determine how the 16S ratio varies under different stages of growth such
as division, metabolic activity, inactivity, and dormancy. This will allow us to better interpret not
only the 16S ratio, but also how changes in 16S rRNA community profiles correlate to physical
measurements of microbial activity such as geochemical rates or enzyme-based assays.
Our results indicated that the active microbial community in the overlying water in marsh
ponds was tightly controlled by diel cycles. The transcriptional activity of autotrophic and
heterotrophic microbial taxa due to diel changes, typically driven by light, has been previously
observed (Winter et al. 2004, Ito et al. 2009, Zinser et al. 2009, Tomasch et al. 2011, Ottesen et
90
al. 2014) and their metabolic activity can be tightly coupled in aquatic systems (Gasol et al.
1998). The changes in the taxonomic composition of the active community observed in our
experiment corresponded with changes in environmental conditions, which are, in turn, driven by
the activity of the microbes themselves. In addition to responding to changes in the environment,
active taxa in our study often appeared to respond to the activity of other taxa in addition to
changes in their environment. For example, the activity of taxa highly abundant in the active
community from the Vibrionales order closely followed the activity of Cyanobacteria (Fig. 4 and
5), suggesting members of the Vibrionales may be involved with either predation or consumption
of cyanobacterial metabolic products, something also documented on seasonal cycles (Turner et
al. 2009). Further, taxa from the order Bacillales were only active when oxygen was at its lowest
(10pm-6am) suggesting that their oxygen requirements as anaerobes restricts their activity. The
tight control of diel cycles on the active community composition in the overlying water
highlights the importance of both abiotic conditions and bacterial interactions in determining
microbial community structure and function.
We demonstrate that taxa from the rare biosphere (<1% total abundance)
disproportionally contributed to the composition of the active community in the overlying water
relative to the sediment (SI Fig. 6). Abiotic conditions within the overlying pond water did not
appear to promote the activity of the most abundant taxa, rather, many of the most abundant taxa
in the active community (Cyanobacteria, Vibrionales, Bacillales) were relatively rare in the total
community. The rare biosphere is often thought of as a reservoir of low-abundance taxa that
provides resilience in the face of environmental perturbations (Sogin et al. 2006, Pedrós-Alió
2012). Rare taxa can display dynamic patterns of occurrence along gradients (Shade and Gilbert
2015) as well as dynamic abundances and activities on seasonal timescales (Campbell et al.
2011, Shade et al. 2014). Shade et al. (2014) recently demonstrated that rare taxa can
significantly contribute to community turnover across numerous ecosystems. Further, rare taxa
can display transitory or persistence patterns within and between ecosystems (Shade and Gilbert
2014), which may allow them to respond to novel abiotic conditions and lead to pulses in their
activity and changes in total community composition (Aanderud et al. 2015). Our results
demonstrate that despite the enhanced activity of rare taxa, there was no net change in the total
community, indicating that either these taxa were not producing biomass through their activity or
not enough time had elapsed to alter total community composition. In addition, the high percent
91
of inactive taxa (Fig. 3), which was likely driven by the dynamic nature of the system,
maintained a reserve of taxa (rare and abundant), allowing for future response to additional
changes within the system.
In conclusion, our results highlight the rapid pace at which microbial communities can
respond to changes in their environment. The dynamic environmental conditions observed in the
overlying water promote changes in active taxa, while the relatively static conditions in the
sediment promote stable communities, highlighting the importance of temporal sampling of
active microbial communities in dynamic environments. Our results suggest that in communities
that are in static environments or that have had time to adapt to a perturbation, the most abundant
taxa are also the most active taxa. However, under dynamic conditions or perhaps during a
transition, there is a disconnect between the presence and activity of taxa leading to high degrees
of inactivity of abundant taxa and an increase in activity of rare taxa. While the pond water
column community and the abiotic conditions changed at a rapid pace, the mechanisms that
control activity of the sediment community likely acted over greater time periods, and those
static conditions promoted a stable active microbial community.
Acknowledgements
We would like to thank Evan Howard, Rachel Stanley, and Amanda Spivak for field
assistance. This work was funded by NSF award DEB1350491 and DEB 1353140) to JLB and
an NSF Research Experience for Undergraduates Award DBI-1359241 to Dr. Rachel Skvirsky.
Additional support was received from the Plum Island LTER (NSF LTER OCE 0423565 and
NSF OCE 1058747). We would also like to thank Jeff Dusenberry for assistance and use of the
UMass Boston high performance computing cluster for bioinformatic analyses. Sequences from
this data set can be found in the Sequence Read Archive under accession number SRR4419663.
Literature Cited
Aanderud ZT, Jones SE, Fierer N, Lennon JT (2015) Resuscitation of the rare biosphere
contributes to pulses of ecosystem activity. Front Microbiol 6: 24.
Aanderud ZT, Vert JC, Lennon JT, Magnusson TW, Breakwell DP, Harker AR (2016) Bacterial
dormancy is more prevalent in freshwater than hypersaline lakes. Front Microbiol 7: 853.
92
Allison SD, Martiny JB (2008) Resistance, resilience, and redundancy in microbial communities.
Proc Natl Acad Sci USA 105: 11512-11519.
Anderson MJ (2001) A new method for non‐parametric multivariate analysis of variance. Austral
Ecol 26: 32-46.
Aronesty, E. (2011) ea-utils: Command-line tools for processing biological sequencing data,
http://code.google.com/p/ea-utils.
Balser TC, Firestone MK (2005). Linking microbial community composition and soil processes
in a California annual grassland and mixed-conifer forest. Biogeochemistry 73: 395-415.
Biddle JF et al., (2008) Metagenomic signatures of the Peru Margin subseafloor biosphere show
a genetically distinct environment. Proc Natl Acad Sci USA 105: 10583–10588.
Blazewicz SJ, Barnard RL, Daly RA, Firestone MK (2013) Evaluating rRNA as an indicator of
microbial activity in environmental communities: limitations and uses. ISME J 7: 2061-2068.
Bowen JL, Ward BB, Morrison HG, Hobbie JE, Valiela I, Deegan LA, Sogin ML (2011)
Microbial community composition in sediments resists perturbation by nutrient enrichment.
ISME J 5: 1540-1548.
Bowen JL, Babbin AR, Kearns PJ, Ward BB (2014) Connecting the dots: linking nitrogen cycle
gene expression to nitrogen fluxes in marine sediment mesocosms. Front Microbiol 5: 429.
Bolhuis H, Stal LJ, (2011) Analysis of bacterial and archaeal diversity in coastal microbial mats
using massive parallel 16S rRNA gene tag sequencing. ISME J 5:1701-1712.
Braman RS, Hendrix SA (1989) Nanogram nitrite and nitrate determination in environmental
and biological-materials by vanadium- (III) reduction with chemi-luminescence detection. Anal
Chem 61: 2715-2718.
Cameron AC, Trivedi PK. (1990) Regression-based tests for overdispersion in the poisson
model. J Econ 46: 347-364.
Campbell BJ, Yu L, Heidelberg JF, Kirchman DL (2011) Activity of abundant and rare bacteria
in a coastal ocean. Proc Natl Acad Sci USA 108: 12776-12781.
Canfield DE, Jørgensen BB, Fossing H, Glud R, Gundersen J, Ramsing NB, Thamdrup B,
Hansen JW, Nielsen LP, Hall PO (1993) Pathways of organic carbon oxidation in three
continental margin sediments. Mar Geol 113: 27-40.Caporaso JG, Bittinger K, Bushman FD,
DeSantis TZ, Andersen GL, Knight R (2010) PyNAST: a flexible tool for aligning sequences to
a template alignment. Bioinformatics 26: 266-267.
93
Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer N, Pena
AG, Goodrich JK, Gordon JI, Huttley GA (2010) QIIME allows analysis of high-throughput
community sequencing data. Nat Meth 7: 335-336.
Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, Fierer N,
Knight R (2011) Global patterns of 16S rRNA diversity at a depth of millions of sequences per
sample. Proc Natl Acad Sci USA 108: 4516-4522.
Day JW, Kemp GP, Reed DJ, Cahoon DR, Boumans RM, Suhayda JM, Gambrell R (2011)
Vegetation death and rapid loss of surface elevation in two contrasting Mississippi delta salt
marshes: The role of sedimentation, autocompaction and sea-level rise. Ecol Eng 37: 229-240.
Deegan LA (1993) Nutrient and energy transport between estuaries and coastal marine
ecosystems by fish migration. Can J Fish Aquat Sci 50: 74-79.
Deutscher J, Francke C, Postma PW (2006) How phosphotransferase system-related protein
phosphorylation regulates carbohydrate metabolism in bacteria. Microbiol Mol Biol Rev 70: 939-
1031.
Edgar RC (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics
26: 2460-2461.
Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R (2011) UCHIME improves sensitivity
and speed of chimera detection. Bioinformatics 27: 2194-2200.
Frossard A, Gerull L, Mutz M, Gessner MO (2012) Disconnect of microbial structure and
function: enzyme activities and bacterial communities in nascent stream corridors. ISME J 6:
680-691.
Garside C (1982) A chemi-luminescent technique for the determination of nanomolar
concentrations of nitrate and nitrite in seawater. Mar Chem 11: 159-167.
Gasol JM, Doval MD, Pinhassi J, Gasol JM, Doval MD, Pinhassi J, Calderón-Paz JI, Guixa-
Boixareu N, Vaqué D, Pedrós-Alió C (1998) Diel variations in bacterial heterotrophic activity
and growth in the northwestern Mediterranean Sea. Mar Ecol Prog Ser 164: 107-124.
Griffiths BS, Philippot L (2013) Insights into the resistance and resilience of the soil microbial
community. FEMS Microbiol Rev 37: 112-129.
Ito H, Mutsuda M, Murayama Y, Tomita J, Hosokawa N, Terauchi K, Sugita C, Sugita M,
Kondo T, Iwasaki H (2009) Cyanobacterial daily life with Kai-based circadian and diurnal
94
genome-wide transcriptional control in Synechococcus elongatus. Proc Natl Acad Sci USA 106:
14168-14173.
Jones SE, Lennon JT (2010) Dormancy contributes to the maintenance of microbial diversity.
Proc Natl Acad Sci USA 107: 5881-5886.
Kearns PJ, Angell JH, Howard EM, Deegan LA, Stanley RH, Bowen JL (2016) Nutrient
enrichment induces dormancy and decreases diversity of active bacteria in salt marsh sediments.
Nat Com 7: 12881.
Kerkhof L, Kemp P (1999) Small ribosomal RNA content in marine Proteobacteria during non-
steady-state growth. FEMS Microbiol Ecol 30: 253-260.
Kirchman DL (2016) Growth rates of microbes in the ocean. Annu Rev Mar Sci 8: 285-309.
Koop-Jakobsen K, Giblin AE (2010) The effect of increased nitrate loading on nitrate reduction
via denitrification and DNRA in salt marsh sediments. Limnol Oceanon 55: 789.
Lennon JT, Jones SE (2011) Microbial seed banks: the ecological and evolutionary implications
of dormancy. Nat Rev Microbiol 9: 119-130.
Lozupone CA, Knight R (2007) Global patterns in bacterial diversity. Proc Natl Acad Sci USA
104: 11436-11440.
Mahé F, Rognes T, Quince C, de Vargas C, Dunthorn M (2014) Swarm: robust and fast
clustering method for amplicon-based studies. PeerJ 2: 1-12.
Mcleod E, Chmura G, Bouillon S, Salm R, Björk M, Duarte C, Lovelock CE, Schlesinger WH,
Silliman BR (2011) A blueprint for blue carbon: toward an improved understanding of the role
of vegetated coastal habitats in sequestering CO. Front Ecol Environ 9: 552-560.
Nemergut DR, Schmidt SK, Fukami T, O'Neill SP, Bilinski TM, Stanish LF, Knelman JE, Darcy
JL, Lynch RC, Wickey P, Ferrenberg S (2013) Patterns and processes of microbial community
assembly. Microbiol Mol Biol Rev 77: 342-356.
Nicholls RJ, Hoozemans FM, Marchand M (1999) Increasing flood risk and wetland losses due
to global sea-level rise: regional and global analyses. Glob Environ Change 9: S69-S87.
Ottesen EA, Young CR, Gifford SM, Eppley JM, Marin R, Schuster SC, Scholin CA, DeLong
EF (2014) Multispecies diel transcriptional oscillations in open ocean heterotrophic bacterial
assemblages. Science 345: 207-212.
Pedrós-Alió C (2012) The rare bacterial biosphere. Annu Rev Mar Sci 4: 449-466.
95
Price MN, Dehal PS, Arkin AP (2010) FastTree 2–approximately maximum-likelihood trees for
large alignments. PloS One 5: e9490.
R Core Team, R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria, (2012).
Redfield AC (1972) Development of a New England salt marsh. Ecol Monogr 42: 201-237.
Reed DJ (2002) Sea-level rise and coastal marsh sustainability: geological and ecological factors
in the Mississippi delta plain. Geomorphology 48: 233-243.
Reed HE, Martiny JB (2013) Microbial composition affects the functioning of estuarine
sediments. ISME J 7: 868-879.
Shade A, Peter H, Allison SD, Baho DL, Berga M, Bürgmann H, Huber DH, Langenheder S,
Lennon JT, Martiny JB, Matulich KL (2012) Fundamentals of microbial community resistance
and resilience. Front Microbiol 3: 417.
Shade A, Gilbert JA (2015) Temporal patterns of rarity provide a more complete view of
microbial diversity. Trend Microbiol 23: 335-340.
Smith KJ, Able KW (2003) Dissolved oxygen dynamics in salt marsh pools and its potential
impacts on fish assemblages. Mar Ecol Prog Ser 258: 223-232.
Sogin ML, Morrison HG, Huber JA, Welch DM, Huse SM, Neal PR, Arrieta JM, Herndl GJ
(2006) Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc Natl
Acad Sci USA 103: 12115-12120.
Stegen JC, Lin X, Konopka AE, Fredrickson JK (2012) Stochastic and deterministic assembly
processes in subsurface microbial communities. The ISME J 6: 1653-1664.
Strickland MS, Lauber C, Fierer N, Bradford MA (2009) Testing the functional significance of
microbial community composition. Ecology 90: 441-451.
Teal J (1962) Energy flow in the salt marsh ecosystem of Georgia. Ecology 43: 614-624.
Tomasch J, Gohl R, Bunk B, Diez MS, Wagner-Döbler I (2011) Transcriptional response of the
photoheterotrophic marine bacterium Dinoroseobacter shibae to changing light regimes. ISME J
5: 1957-1968.
Turner JW, Good B, Cole D, Lipp EK (2009) Plankton composition and environmental factors
contribute to Vibrio seasonality. ISME J 3: 1082-1092.
Winter C, Herndl GJ, Weinbauer MG (2004). Diel cycles in viral infection of bacterioplankton in
the North Sea. Aquat Microb Ecol 35: 207-216.
96
Zinser ER, Lindell D, Johnson ZI, Futschik ME, Steglich C, Coleman ML, Wright MA, Rector
T, Steen R, McNulty N, Thompson LR (2009) Choreography of the transcriptome,
photophysiology, and cell cycle of a minimal photoautotroph, Prochlorococcus. PLoS One 4:
e5135.
97
Figures
Figure 1- Figure 1- Environmental parameters in the sediment pore water (black) and overlying
water (grey) from ammonium (A), Phosphate (B), salinity (C), pH (D), C:N ratio (E), dissolved
oxygen (F), and temperature (G). Grey bars are night time hours. Points are the average of all
three sites and error bars indicate the standard error of the mean.
98
Figure 2- Principal coordinates analysis based on weighted UniFrac similarity for the sediment
(A) and the overlying water (B).
99
Figure 3- Box and whisker plot of the percentage of inactive taxa in the sediment (A) and water
(B) communities. Taxa are defined inactive if their 16S rRNA:16S rRNA gene ratio is less than
1. Boxes represent 25-75% quartiles, and the solid black line is the median value. Categories in
(B) are significant based on a Tukey HSD correcting for multiple comparisons and no significant
differences occurred in (A).
100
Figure 4- Stacked bar plot of the top 25 most abundant active microbial orders in the sediment
(A) and water (B). The category “other” is the sum of the remaining microbial orders.
101
Figure 5- Heat map showing the sum of the 16S rRNA: 16S rRNA gene ratio for 28 most active
microbial orders in the water column.
102
Figure 6- Plot of the relative abundance of the top 1000 OTUs (comprising >90% of the data) in
16S rRNA relative to the 16S rRNA gene for sediment (A) and water communities (B). Points
are colored by the time point they are most abundant in. Black lines in (A) and (B) are 1:1 lines.
103
Figure 7-plot of the 16S rRNA gene rank of the top 380 OTUs (comprising ~90% of the data)
versus the relative abundance in 16S rRNA for sediment (A) and overlying water communities
(B). Inset of (B) highlights the top 60 OTUs in the water. Vertical red lines indicate taxa with an
abundance in the total community <1% and vertical blue lines are taxa with an abundance in the
total community <0.001%.
104
Chapter 3: Supplemental Material
SI Figure 1- Quantitative PCR (qPCR) results depicting the ratio of 16S rRNA:16S rRNA gene
for sediment (black) and water (grey) communities. Points are the mean of three sites and the
error bars are standard error of the mean.
105
SI Figure 2- Shannon Diversity values for sediment (A) and water (B) communities. Boxes
represent 25-75% quartiles, the solid black line is the median value, and dots are outliers.
Categories in (B) are significant (p<0.001) based on a Tukey HSD correction for multiple
comparisons.
106
SI Figure 3- Stacked bar plot of the top 25 most abundant microbial orders from the sediment
16S rRNA gene. The category other is the sum of the remaining microbial orders. Black bars are
night time hours.
107
SI Figure 4- Stacked bar plot of the top 25 most abundant microbial orders from the water 16S
rRNA gene. The category other is the sum of the remaining microbial orders. Black bars are
night time hours.
108
SI Figure 5- Heat map showing the sum of the 16S rRNA: 16S rRNA gene ratio for 25 most
active microbial orders in the sediment.