Subscriber access provided by UNIV OF LOUISIANA
Environmental Processes
Combined deterministic and stochastic processes control microbial succession in replicate granular biofilm reactors Raquel Liébana, Oskar Modin, Frank Persson, Enikö Szabó, Malte Hermansson, and Britt-Marie Wilen Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b06669 • Publication Date (Web): 10 Apr 2019 Downloaded from http://pubs.acs.org on April 10, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 39
Environmental Science & Technology
1
Combined deterministic and stochastic processes control microbial succession in replicate
2
granular biofilm reactors
3
Raquel Liébana a*, Oskar Modin a, Frank Persson a, Enikö Szabó a, Malte Hermansson b and Britt-
4
Marie Wilén a
5 6
a
7
Chalmers University of Technology, Gothenburg, Sweden
8
b
9
Sweden
Division of Water Environment Technology, Department of Architecture and Civil Engineering,
Department of Chemistry and Molecular Biology, University of Gothenburg, Gothenburg,
10
*Corresponding author: Raquel Liébana; Postal address: Division of Water Environment
11
Technology, Department of Architecture and Civil Engineering, Chalmers University of
12
Technology, Sven Hultins gata 8, SE- 412 96 Gothenburg, Sweden; Telephone number: +46 031-
13
772-1912, Email address:
[email protected] 14 15 16 17 1
ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 39
18 19
ABSTRACT
20
Granular sludge is an efficient and compact biofilm process for wastewater treatment. However,
21
the ecological factors involved in microbial community assembly during the granular biofilm
22
formation are poorly understood and little is known about the reproducibility of the process.
23
Here, three replicate bioreactors were used to investigate microbial succession during the
24
formation of granular biofilms. We identified three successional phases. During the initial phase,
25
the successional turnover was high and α-diversity decreased as a result of the selection of taxa
26
adapted to grow on acetate and form aggregates. Despite these dynamic changes, the microbial
27
communities in the replicate reactors were similar. The second successional phase occurred when
28
the settling time was rapidly decreased to selectively retain granules in the reactors. The influence
29
of stochasticity on succession increased and new niches were created as granules emerged,
30
resulting in temporarily increased α-diversity. The third successional phase occurred when the
31
settling time was kept stable and granules dominated the biomass. Turnover was low and
32
selection resulted in the same abundant taxa in the reactors, but drift, which mostly affected low-
33
abundant community members, caused the community in one reactor to diverge from the other
34
two. Even so, performance was stable and similar between reactors.
35 36
2
ACS Paragon Plus Environment
Page 3 of 39
Environmental Science & Technology
37 38
TABLE OF CONTENTS (TOC)/ABSTRACT ART
39 40 41
INTRODUCTION
42
The structure of microbial communities results from complex and dynamic ecological
43
mechanisms 1, which traditionally have been grouped into deterministic and stochastic factors 2.
44
Environmental conditions, species interactions and species traits are considered deterministic 3,
45
4,
46
deterministic and stochastic factors can be framed into four fundamental processes: selection,
47
dispersal, diversification and drift
48
community structure according to the environmental conditions, differences in fitness between
49
individuals and microbial interactions. Dispersal refers to the movement and establishment of
50
microorganisms among communities, which can be both deterministic and stochastic.
51
Diversification refers mainly to stochastic factors, which generate genetic variation. Drift refers
whereas random events, such as birth and death, are considered stochastic 5. These
6, 7.
Selection refers to deterministic factors that modify the
3
ACS Paragon Plus Environment
Environmental Science & Technology
Page 4 of 39
6, 7.
52
to stochastic changes as a result of birth, death and reproduction
53
simultaneously in natural ecosystems and their influences vary in time and space 2, 7, therefore it
54
is challenging to determine their contribution to microbial community assembly. The study of the
55
biodiversity within and between microbial communities can help us to understand the underlying
56
ecological processes 7. Studying the compositional differences (turnover) between two different
57
communities (β-diversity) provides a link between biodiversity at the local scale or at
58
instantaneous sampling (within-sample or α-diversity) and at the regional scale or long time scale
59
(overall diversity or γ-diversity) 8.
60
The importance ascribed to deterministic and stochastic factors during microbial community
61
assembly in wastewater bioreactors vary. Some studies report similar microbial communities in
62
replicate reactors, such as membrane bioreactors 9, anaerobic digesters 10-12, microbial fuel cells
63
13,
64
environmental conditions. On the contrary, diverging communities and functions have been
65
reported in replicate microbial electrolysis cell reactors
66
anaerobic digesters 20, 21, due to the roles of stochastic factors.
67
Wastewater treatment bioreactors are engineered to select for functional groups needed for
68
water purification, e.g. nitrifiers, denitrifiers, and phosphate accumulating organisms. Moreover,
69
the reactor operational conditions are adjusted to cultivate the microorganisms in the desired
70
aggregate modes, either as flocs in activated sludge or as biofilms on different substrata
71
Free-floating spherical biofilms, so called granules, combine many of the properties of these two
biofilters
14, 15
and sequencing batch reactors
16, 17,
These four processes act
due to selection caused by the reactor
18,
sequencing batch reactors
19
and
22, 23.
4
ACS Paragon Plus Environment
Page 5 of 39
Environmental Science & Technology
72
growth modes. The granules are compact and spherical suspended biofilms with a diameter of
73
approximately 1-3 mm obtained at defined reactor conditions of large shear forces created by a
74
high up-flow air velocity, typically higher than 1 cm s−1, large temporal variation of electron donors
75
and -acceptors causing feast-famine operation, and short settling time to select for well settling
76
biomass 24. Granules are stratified in an aerobic outer layer and an anoxic substrate-rich interior
77
as a consequence of oxygen and substrate gradients inside the granule where heterotrophic-,
78
nitrifying-, denitrifying-, phosphorous-accumulating- and glycogen-accumulating bacteria can
79
coexist
80
simultaneously, which renders highly compact and energy efficient wastewater treatment
81
Despite the well-established methods for granule cultivation, the ecological processes
82
underpinning the microbial community assembly during granulation are poorly understood
83
Granulation is believed to occur as a response to specific selection pressures created in the
84
reactor, which causes the bacteria to switch from a planktonic to an aggregate growth mode 28.
85
High shear forces and feast-famine alternation stimulate the bacteria to increase the production
86
of extracellular polymeric substances (EPS) and increases the bacterial cell surface
87
hydrophobicity, accelerating the microbial aggregation 29. High wash out dynamics would act as
88
an accelerant of granulation by the physical selection of bigger particles 25.
89
Over the last decade, many studies have been conducted in granular sludge analysing the
90
microbial dynamics descriptively. However, there is a need to understand the factors shaping the
91
granular microbial community. For this, laboratory scale studies that enable controlled
92
environmental conditions and tests of reproducibility are valuable test-benches 1. Laboratory
25.
This permits granular sludge to remove carbon, nitrogen and phosphorus 26.
27.
5
ACS Paragon Plus Environment
Environmental Science & Technology
Page 6 of 39
93
experiments in granular sludge reactors, and generally in the field of wastewater treatment, are
94
typically performed in one reactor because of practical reasons. Hence, conclusions regarding
95
reactor performance and microbial community structure are drawn from single reactors operated
96
at different conditions 30-33. For example, in a previous study we showed considerable divergence
97
of microbial communities in three sequencing batch reactors, each fed with different carbon
98
concentrations, during granulation 30. It is, however, unclear how reliable such conclusions are,
99
especially for the complex processes underlying microbial community assembly. It is therefore
100
necessary to assess the reproducibility of these systems.
101
Here, to better understand the microbial community assembly of granule formation, we have
102
investigated the succession during the transition from floccular to granular sludge in replicate
103
reactors. Reactor functions, granule development, successional patterns of taxa, microbial
104
diversity and reproducibility were assessed. We used null models to determine the ecological
105
factors behind taxonomic- and phylogenetic turnover to assess community assembly. We
106
hypothesized that similar communities would develop in the three replicate reactors if
107
deterministic factors were dominating, whereas the opposite would support the importance of
108
stochastic factors.
109
MATERIAL AND METHODS
110
Reactors and analytical methods
6
ACS Paragon Plus Environment
Page 7 of 39
Environmental Science & Technology
111
Three replicate sequencing batch reactors (R1, R2 and R3) with a working volume of 3 L,
112
previously described in detail
113
inoculated with fresh activated sludge for a total addition of 9 g of biomass to each reactor. A
114
mineral medium with acetate as the single source of organic carbon was used with a chemical
115
oxygen demand (COD) of 3 kg COD m-3d-1 and a nitrogen load of 0.75 kg NH4-N m-3d-1 (see Text S1
116
and Figure S1 for detailed information). Concentrations in the effluent of total organic carbon and
117
total nitrogen were measured with a TOC-TN analyser (TOC-V, Shimadzu, Japan). Total and
118
volatile suspended solids in the reactor and in the effluent were measured according to standard
119
methods
120
microscope (Olympus Sverige AB, Solna, Sweden), using the CellSens (Olympus) software to
121
measure the diameter of 10 random flocs/granules.
122
DNA extraction, amplification and sequencing
123
The V4 region of the 16S rRNA gene was amplified using indexed 515'F and 806R primers 35 and
124
sequenced using the MiSeq platform (Illumina). For detailed information, see Text S1. To increase
125
the confidence in the obtained results, a second MiSeq sequencing run was performed on the
126
same DNA pool.
127
Sequence processing and data analysis
128
Raw sequence reads were processed following the UNOISE pipeline 36, 37 with USEARCH v.10 38,
129
taxonomically classified with the SINTAX algorithm 39 based on the MiDAS database v.2.1 40 and
34.
30,
were operated identically for 35 days. The reactors were
Sludge particle size assessments were performed with an Olympus BX60 light
7
ACS Paragon Plus Environment
Environmental Science & Technology
Page 8 of 39
130
analysed in R version 3.4.1 (http://www.r-project.org). For comparison, raw sequence reads were
131
also processed following the DADA2 pipeline 41. An approximately maximum-likelihood tree was
132
generated using the R package DECIPHER 42 and the FastTree 2 software 43. Raw sequence reads
133
from both MiSeq sequencing runs were processed following the same procedure. The dataset
134
was rarefied, subsampling each sample to 17395 sequences. Basic R functions were used to
135
perform Wilcoxon signed-rank tests and to calculate Pearson correlation coefficients. NMDS
136
ordination and heatmaps were created using the R package ampvis
137
permutations) and Procrustes tests using the protest function (999 permutations) were
138
performed using the R package vegan 45. Analysis of pairwise multivariate permutational analysis
139
of variances (PERMANOVA) was conducted using adonis from the package pairwiseAdonis 46. The
140
ANCOM method 47 was implemented in scikit-bio 0.5.5. Time-decay rates were calculated as in
141
Shade, et al. 48. For detailed information on sequence processing and data analysis, see Text S1.
142
Taxonomic Hill numbers were used to calculate α-diversity, qTD 49. Phylogenetic diversity, qPD,
143
which take the sequence dissimilarity into account, was also calculated 50. The same calculation
144
framework was used to calculate β-diversity and was converted into dissimilarity indices (qβdis)
145
constrained between 0 (two identical samples) and 1 (two samples with no shared OTUs) 50, 51.
146
The parameters qβdisTD and qβdisPD are based on the taxonomic and the phylogenetic β-diversity
147
values, respectively (see TextS1 for detailed information). Correlations between series of α-
148
diversity data were investigated using Kendall’s rank correlation coefficient (tau), which was
149
calculated with the Scipy package in Python 52 (see Text S1).
44.
CAP, Mantel tests (999
8
ACS Paragon Plus Environment
Page 9 of 39
Environmental Science & Technology
150
Raw sequences were deposited in the NCBI Sequence Read Archive (SRA), BioProjectID
151
PRJNA472243, accession number SRP148672. Python scripts used to calculate diversity indices
152
based on Hill numbers are freely available at https://github.com/omvatten/qDiv.
153
Null model analysis
154
Taxonomic turnover was estimated with Raup-Crick measures based on Bray-Curtis dissimilarities
155
(RCbray) as by Stegen, et al. 53. The RCbray index values range between -1 and 1. A negative value
156
means that the two communities are more similar than expected by chance whereas a positive
157
value means that they are more dissimilar. Values > |0.95| were considered statistically
158
significant. Values < |0.95| indicates that the taxonomic turnover between the community pair
159
are not different from the null expectation and therefore, influenced by stochastic factors 2.
160
Phylogenetic turnover was estimated with β nearest-taxon index (βNTI), as previously described
161
53. A negative βNTI value means that the samples are more phylogenetically similar than expected
162
by chance and a positive value means that they are more phylogenetically distant from each
163
other. Pairwise comparisons with βNTI > |2| were considered statistically significant. Values < |2|
164
were not significantly different from the null expectation, which indicate that stochastic factors
165
influenced the phylogenetic turnover 2. For detailed information see Text S1.
166
RESULTS
167
Granulation of the sludge and reactor performance
9
ACS Paragon Plus Environment
Environmental Science & Technology
Page 10 of 39
168
Around day 7, granules started to emerge (Figure S2) and the mean particle sizes of the reactor
169
biomass increased similarly in the three reactors (Figure S3). On day 25, the granulated biomass
170
dominated and at the end of the experiment (day 35), the granules had an average diameter of
171
approximately 1.5 mm. The three reactors were reproducible in terms of effluent concentrations
172
of suspended solids (total and volatile), organic carbon and nitrogen (Figure 1) as indicated by
173
paired sample Wilcoxon signed-rank tests (Table S1), but the reactor concentrations of volatile
174
and total suspended solids were significantly lower (p
