Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Linking microbial community, environmental variables and methanogenesis in anaerobic biogas digesters of chemically enhanced primary treatment sludge Feng Ju, Frankie Lau, and Tong Zhang Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b06344 • Publication Date (Web): 27 Feb 2017 Downloaded from http://pubs.acs.org on February 28, 2017
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 free 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 accessible to all readers and 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.
Environmental Science & Technology 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 40
Environmental Science & Technology
Linking microbial community, environmental variables and methanogenesis in anaerobic biogas digesters of chemically enhanced primary treatment sludge Authors: Feng Ju 1, Frankie Lau 2, Tong Zhang 1* Author affiliation:
1
Environmental Biotechnology Lab, The University of Hong
Kong SAR,China; 2Drainage Services Department, The Government of the Hong Kong Special Administrative Region, Hong Kong, China. *Corresponding author: Dr. Tong Zhang (Professor) Address: Environmental Biotechnology Lab, The University of Hong Kong, Pokfulam Road, Hong Kong Tel: 852-28591968 (lab), 28578551 (office) Fax: 852-25595337 E-mail:
[email protected] 1
ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 40
1
Abstract
2
Understanding the influences of biotic and abiotic factors on microbial community structure
3
and methanogenesis are important for its engineering and ecological significance. In this study,
4
four biogas digesters were supplied with the same inoculum and feeding sludge, but operated
5
at different sludge retention time (7 to 16 days) and organic loading rates for 90 days to
6
determine the relative influence of biotic and environmental factors on the microbial
7
community assembly and methanogenic performance. Despite different operational
8
parameters, all digester communities were dominated by Bacteroidales, Clostridiales and
9
Thermotogales, and followed the same trend of population dynamics over time. Network and
10
multivariate analyses suggest that deterministic factors, including microbial competition
11
(involving Bacteroidales spp.), niche differentiation (e.g., within Clostridiales spp.), and
12
periodic microbial immigration (from feed sludge), are the key drivers of microbial
13
community assembly and dynamics. A yet-to-be-cultured phylotype of Bacteroidales
14
(GenBank ID: GU389558.1) is implicated as a strong competitor for carbohydrates. Moreover,
15
biogas-producing rate and methane content were significantly related with the abundances of
16
functional populations rather than any operational or physicochemical parameter, revealing
17
microbiological mediation of methanogenesis. Combined, this study enriches our
18
understandings of biological and environmental drivers of microbial community assembly and
19
performance in anaerobic digesters.
20
Keywords: Anaerobic digesters; Methanogenesis; Microbial community; Population 2
ACS Paragon Plus Environment
Page 3 of 40
Environmental Science & Technology
21
dynamics; Network analysis
22
Introduction
23
Anaerobic digestion (AD) is a microbially mediated biotechnology widely used for renewable
24
energy production and waste management 1, 2. Environmental variables, including operational
25
parameters (e.g., sludge retention time) and physicochemical conditions (e.g., volatile fatty
26
acids (VFAs), ammonium nitrogen (NH4-N)), are known to affect AD process, including
27
formation of intermediates, process stability, biogas yield, biodegradation efficiency, and 3, 4, 5.
28
However, because of the lack of dense monitoring of coupled environmental and
29
microbiological data from the same anaerobic digesters, biotic and abiotic factors that govern
30
microbial community structure and their linkage with methane production are poorly
31
understood3, 6, 7.
32
Recent application of high-throughput sequencing technologies provides a comprehensive,
33
qualitative and quantitative measurement of diverse environmental microbiomes, facilitating
34
the exploration of the driving forces that structure microbial communities in marine water 8,
35
lake 9, soil
36
interpret patterns and processes of microbial community assembly
37
highlights an importance of deterministic processes, such as microbial interactions (e.g.,
38
syntrophy and competition), niche differentiation, substrate availability and operational
39
conditions 6, 11. On the contrary, neutral theory, as a null hypothesis, disregards the differences
40
among species in response to ecological conditions and only considers stochastic processes,
10-12
, and bioreactors
6, 13-15
. Both niche-based and neutral theories are used to 16
. Niche-based theory
3
ACS Paragon Plus Environment
Environmental Science & Technology
Page 4 of 40
15, 17
41
such as birth, death, dispersal, and colonization and immigration
42
the relative role of niche and neutral forces in structuring microbiome are controversial. Some
43
researchers argue that microbiome assembly in lake water, soil, and activated sludge is driven
44
by deterministic niche-based processes based on co-occurrence analysis
45
reactor experiments 6; others that neutral models support stochastic processes for interpreting
46
microbiome assembly in marine water 8, activated sludge
47
that niche-based and stochastic-neutral processes are both important for structuring
48
microbiomes of soil and bioreactor 10.
49
Bioreactors are ideal systems to study patterns of microbial community assembly and link
50
them with system performance, because microorganisms are cultivated in elaborately
51
controlled and regularly monitored systems
52
previously reported in quintuplicate denitrifying bioreactors
53
membrane reactors
54
suggests that these highly predictable microbial systems are not guided by stochastic
55
processes. However, replicate operational conditions, such as sludge retention time (SRT) and
56
organic loading rate (OLR), and highly similar profiles of chemicals (e.g., VFAs and NH4-N)
57
in these replicated bioreactors make it impossible to resolve the influence of different
58
operational parameters on the microbial community assembly and performance. In fact,
59
operational parameters including SRT and NH4-N have long been implicated to affect
60
microbial community structure and methane production
15
. Recent viewpoints on
9, 13
or replicated
and bioreactors 18; and yet more
6, 13, 14
. The reproducible community dynamics
20
19
, quadruplicate nitrifying
, and triplicate anaerobic cellulose-degrading reactors
6
congruously
3, 4
. Therefore, the examination of
4
ACS Paragon Plus Environment
Page 5 of 40
Environmental Science & Technology
61
anaerobic digesters under different operational conditions is essential for checking the
62
robustness of deterministic community assembly and dynamics observed in previous replicate
63
bioreactors, examining whether different operational conditions lead to stochastic or random
64
community assemblages, and determining the influences of biotic and abiotic (environmental)
65
factors in shaping community structure and performance.
66
Here, four anaerobic digesters were supplied with the same seeding and feeding sludge and
67
operated at different SRTs and OLRs. Temporal dynamics in the microbial communities and
68
environmental conditions was closely followed over 90 days and linked to methane
69
production. This work was conducted to show the relationships among microbial community,
70
environmental variables and methane production, to check whether microbial community
71
assembly and dynamics were deterministic or stochastic, and to compare the relative
72
influence of environmental variables (operational and physicochemical) and biotic factors
73
(e.g., microbial interactions) in shaping microbial community assembly and performance.
74
Materials and Methods
75
Digester operation and sample collection
76
Four 1-liter continuously stirring anaerobic digesters of chemically enhanced primary
77
treatment (CEPT) sludge, designated as R1, R2, R3 and R4, were operated at 35°C (heated in
78
water bath) but under different SRTs of 7, 9, 12 and 16 days, corresponding to levels of OLR
79
between 2.40 and 3.78, 1.86 and 2.93, 1.40 and 2.21, and 1.05 and 1.66 kg-VS/m3/d, 5
ACS Paragon Plus Environment
Environmental Science & Technology
Page 6 of 40
80
respectively (Table S1). The setup of retention time below and above 10 days, which is
81
deemed as the minimum retention time needed to prevent washout of slow-growing
82
methanogens by some researchers
83
recommended minimum SRT on the methanogenic performance and microbial structure.
84
More details on digester operation are available in the Supporting Method S1. For
85
microbiological analysis, the seed sludge (in triplicates), feed sludge (Day 1), and 48 digester
86
samples (collected at 12 time points; Figure 1a) were fixed in 50% ethanol and stored at
87
-20°C until further processing.
88
Digester performance monitoring
89
Biogas, liquid, and sludge samples were collected for chemical analysis every 2 or 4 days
90
from the four anaerobic digesters, as described in Supporting Method S2. The digesters
91
showed reproducible performance in biogas production and volatile solid reduction over 90
92
days (Figure S1).
93
DNA extraction, sequencing and bioinformatics analysis
94
16S rRNA genes were amplified with the universal primer set F515/R806 11, 21 from genomic
95
DNAs extracted using FastDNA SPIN Kit for Soil. The pooled PCR products were purified
96
and prepared for paired-end sequencing (2×250 bp) on Illumina Miseq. The generated data
97
were pretreated following mothur Miseq SOP
98
QIIME (v 1.8.0) for open-reference operational taxonomic units (OTUs) picking and core
99
diversity analyses 23. The molecular and bioinformatics procedures are described in details in
4, 6
, allows us to test the effects of a span around the
22
, and the clean data were imported into
6
ACS Paragon Plus Environment
Page 7 of 40
Environmental Science & Technology
100
Supporting Method S3. The sequence data are deposited into the NCBI’s Sequence Read
101
Archive Database with accession numbers of SRR3104406, SRR3104407, SRR3104408,
102
SRR3104410 and SRR3104411.
103
Statistical and network analyses
104
The statistical analyses including principal coordinates analysis (PCoA), procrustes analysis,
105
canonical correspondence analysis (CCA), correlation analysis, student’s t-test, Wilcoxon
106
signed-rank test, Kruskal-Wallis tests and BIOENV analysis were performed using R
107
packages: stats
108
and co-exclusion species-species associations (SSA) were explored by calculating all pairwise
109
Spearman’s rank coefficients (ρ) among 0.97-OTUs that occurred in at least 50% of samples
110
and represented by at least five sequences on average across all samples. Details on the
111
statistical and network analyses were described in the Supporting Method S4.
112
Results
113
Digester performance and operational parameter
114
Four anaerobic digesters were operated to investigate the effects of different SRTs (7-16 days)
115
and OLRs (1.05-3.78 kg-VS/m3/d) on the biogas production and microbial community
116
structure over 90 days. To facilitate comparison of temporal succession of microbial
117
community, three operational stages with similar time lengths were defined: Stage I—day
118
10-38, Stage II—day 39-66, and Stage III—day 67-90 (Figure 1).
24
, vegan
25
, MASS
26
and nortest 27, unless stated otherwise. Co-occurrence
7
ACS Paragon Plus Environment
Environmental Science & Technology
Page 8 of 40
119
The four anaerobic digesters achieved good methane-producing rate (RM) (0.49-0.66 L/L/d;
120
Figure 1b) and methane content (CH4%; 64-67%, Table S1) at Day 12, indicating rapid
121
adaptation of the inoculated microorganisms to the feedings of sewage sludge. From day 10 to
122
18, R1 and R2 (shorter SRTs of 7 and 9 days) had higher RM than R3 and R4 (higher SRTs of
123
12 and 16 days), probably because of the higher OLRs to the former over the latter digesters
124
(Figure 1a, inset). However, accumulation of VFAs (mainly acetate and propionate) were
125
observed in R1 during stage I (Figure 1c and Table S2) and RM dramatically decreased on Day
126
34 (Figure 1b), closely followed by the peak of VFAs at Day 38 (Table S2). Stage II was
127
characterized by OLR decrease from day 42 to 50 and OLR increase from day 54 to 66. This
128
accompanied with the decrease of VFA concentrations in R1 (33-128 mg/L), as well as
129
fluctuations of RM and pH in all digesters (Figure 1b and 1d). During Stage III, the digester
130
OLRs were markedly increased on Day 66 (by 38%), followed by dramatic increases of RM in
131
all digesters expect for R1. Two-sample statistical tests show that RM from day 10 to 90 were
132
significantly (P-values < 2.3×10-4) different between R2 (0.95±0.14 L/L/d) and R3
133
(0.73±0.20), between R3 and R4 (0.59±0.10), and between R2 and R1 (0.54±0.14). In
134
contrast, there was no significant (P-values > 0.07) difference in CH4% between any two
135
digesters.
136
Associations among environmental parameters, microbial alpha-diversity and biogas
137
parameters
138
Pearson’s linear correlations and/or Spearman’s rank correlations (Table S4) show that RM 8
ACS Paragon Plus Environment
Page 9 of 40
Environmental Science & Technology
139
was negatively correlated with the concentration of organic parameters, including acetic acid
140
(HAc), propionic acid (HPr) and VFAs (-(0.39-0.48), Table S4-1), as well as microbial
141
alpha-diversity metrics, such as Shannon’s and Simpson’s diversity indexes (-(0.38-0.40),
142
Table S4-2). Moreover, all alpha-diversity metrics were negatively correlated with operating
143
time (i.e., Day; -(0.61-0.84)), inorganic carbon (IC, -(0.37-0.65)), and NH4-N (-(0.41-0.64)
144
(Table S4-1). This is in agreement with the results of two-sample statistical tests that the
145
richness, evenness and diversity metrics of 0.97-OTUs significantly decreased (P-values ≤
146
0.05) with operating time from Stage I to III (Table 1).
147
Microbial population dynamics in SRT-differentiated digesters
148
The seed sludge was mainly composed of Bacteroidia (20.9%), Gammaproteobacteria
149
(12.7%) and Betaproteobacteria (8.5%). All the digester microbial communities rapidly
150
shifted away from the seed community (Figure 2, 0.97-OTU level; Figure S2, class level) and
151
markedly decreased in richness (i.e., observed species) and Shannon’s H diversity of OTUs
152
after startup (Figure 2, diversity heatmap). PCoA biplots based on UniFrac distances (Figure 3)
153
and Bray Curtis dissimilarities between 0.97-OTUs (Figure S3a) show that both microbiome
154
membership (i.e., presence or absence of rare phylotypes) and abundance were much more
155
different between operating stages (Figure 3c & 3d) than between digesters (Figure 3a & 3b).
156
Procrustes analysis on the principal coordinates of unweighted UniFrac and weighted UniFrac
157
distances between samples over 90 days shows a significant and high correlation (0.85,
158
P-value = 0.001) between microbial community structure with and without considering 9
ACS Paragon Plus Environment
Environmental Science & Technology
Page 10 of 40
159
membership, and highly similar community dynamics among the four digesters (Figure S3b).
160
BIOENV analysis supports that operating time best explained the variations of microbial
161
community abundance and membership (rho = 0.518-0.520, weighted UniFrac distance,
162
Table 2). However, compared with only considering operating time, the incorporation of SRT
163
slightly improved interpretation of the unweighted (but not weighted) UniFrac distances
164
between the samples (rho from 0.369-0.413 to 0.442-0.481, Table 2), revealing that SRT
165
slightly determines microbial community membership. PCoA bi-plots highlighted the
166
influence of SRT on the community membership at Stage I. For example, the community
167
membership (Figure 3d) changed more apparently than abundance (Figure 3c) for all
168
digesters.
169
Despite the differences in SRT and operating time, the microbial communities in all digester
170
samples (n = 48) were composed of Bacteroidia (42.8±13.1%), Clostridia (23.0±6.2%),
171
Thermotogae (8.7±7.0%), Bacilli (3.0±1.3%), Anaerolineae (2.7±1.2%), Actinobacteria
172
(2.1±1.0%) and other populations (17.8±5.2%) (Figure S2, clusters a and c1). Moreover, the
173
overall profiles of microbial abundance and diversity metrics in the four digesters were
174
similar over time (Figure S2) and the microbiome succession followed deterministic trends
175
over 90 days (Figure 3 and Figure S3b). The 50 most abundant 0.97-OTUs in the four
176
digesters over 90 days (i.e., labeled from s1 to s50; Table S5) accounted for total relative
177
abundance of 71.1(±6.8)%. Among them, 38 OTUs had significant different (P-values ≤ 0.05)
178
relative abundances between the seed sludge and different operating stages (29 OTUs, taxon
10
ACS Paragon Plus Environment
Page 11 of 40
Environmental Science & Technology
179
names in red) or SRTs (10 OTUs, taxa names in green), as shown by Kruskal-Wallis tests
180
(Figure 2, P-value heatmap).
181
Environment-species and species-performance associations
182
The relationships among environmental (i.e., operational and physicochemical) parameters,
183
abundance of 0.97-OTUs, and methanogenic performance were explored by CCA based on
184
data from all 48 digester samples. Overall, the associations between environmental and
185
performance parameters, as shown by CCA plot, were in agreement with the results of
186
correlation analysis (Table S4). CCA additionally shows that Bacteroidales OTUs including
187
s1, s13 and s49 (red nodes, Figure 4) were positively correlated with both methane-producing
188
rate (RM) and biogas-producing rate (RG) whereas negatively with CH4%. The most abundant
189
OTU, s1 (GenBank ID: GU389558.1, Greengenes ID: 837605, Figure S6), was significantly
190
enriched (P-value ≤ 0.01) in all digesters, from 1.1% in the seed sludge to average relative
191
abundances of 15.1% 28.0% and 45.6% at Stages I (n=16), II (n=16) and III (n=16),
192
respectively (Figure 2, cluster a). In contrast, the organic parameters, including OLR, DOC,
193
VFAs, HPr and HAc, were positively correlated a group of OTUs (blue nodes, Figure 4),
194
including four Clostridium OTUs (s3, s4, s19 and s36), four Syntrophomonas OTUs (s5, s12,
195
s18 and s29), three Bacteroidales OTUs (s10, s15 and s46), three Verruco-5 OTUs (of
196
Verrucomicrobia; s28, s38 and s48), two T78 clade OTUs (of Chloroflexi, s9 and s23) and one
197
Fibrobacter succinogenes OTU (s17). The most abundant methanogen, s30, (Greengenes
198
taxonomy ID: 103247) was correlated positively with HAc, whereas negatively with SRT 11
ACS Paragon Plus Environment
Environmental Science & Technology
Page 12 of 40
199
(Figure 4). This Methanosaeta species had significantly different relative abundances at
200
different SRTs (Kruskal-Wallis test P-value ≤ 0.001; Figure 2, P-value heatmap). Likewise,
201
methanogenic populations of class Methanomicrobia and genus Methanosaeta were the most
202
enriched at the SRT of 7 days, followed by SRTs of 9, 12 and 16 days, respectively (Figure
203
S4).
204
Species-species associations
205
The co-occurrence (i.e., positive) and co-exclusion (i.e., negative) associations between
206
species (i.e., 0.97-OTUs) over time were explored using network analysis of 48 digester
207
samples collected from Day 10 to 90. The resulting positive SSA network consisted of 101
208
nodes (i.e., 0.97-OTUs) and 317 edges (i.e., correlation; Figure 5a), while the negative SSA
209
network included 60 nodes and 131 edges (Figure 5b). Overall, 52% and 48% of OTUs in the
210
positive and negative SSA networks belonged to Bacteroidetes and Firmicutes, respectively.
211
Furthermore, 20% and 40% of all the 131 negative SSA involved the most abundant,
212
Bacteroidales s1 and other 15 Bacteroidales OTUs, respectively. Notably, the average
213
correlation coefficients of negative SSA between s1 and other OTUs (-0.71; see Figure 5d for
214
examples) were significant higher (P-values < 0.0003, Figure 5c) than those between other 15
215
Bacteroidales OTUs and between 44 non-Bacteroidales OTUs.
216
Topological analysis shows that the positive SSA network had a clustering coefficient of 0.47
217
and a modularity index of 0.49. Network partitioning divided the network into seven modules
12
ACS Paragon Plus Environment
Page 13 of 40
Environmental Science & Technology
218
(Figure S5a). Module I was aggregated by 25 densely connected OTUs with peak abundances
219
at Stage I (see green nodes in Figure S5a against Figure S5b). However, these OTUs had
220
sharply decreasing abundance with operating time. This was accompanied with peak
221
abundance of many other OTUs (mainly in Modules II and V, Figure S5a) at Stage II (39
222
OTUs; red nodes) or Stage III (14 OTUs, purple nodes; Figure S5b). Different from Module I
223
where OTUs of Actinobacteria, Tenericutes and WWE1 were prevalent (see Figure S5a
224
against 5a), Module II mainly harbored OTUs of Bacteroidetes (10 OTUs), Proteobacteria (7
225
OTUs), Verrucomicrobia (3 OTUs), and Firmicutes (3 OTUs).
226
Further statistical analysis reveals non-random (deterministic) co-exclusion patterns between
227
OTUs from different orders (i.e., inter-orders) or from the same order (i.e., intra-order). Firstly,
228
OTUs from order Bacteroidales tended to co-exclude with OTUs of other bacterial orders (i.e.,
229
Actinomycetales, Cloacamonales, Mollicutes, Anaerolineales and Synergistales) much more
230
(23%) than expected when considering observed order frequencies and random association
231
(6%, Table S6). Moreover, the observed incidence of intra-order co-exclusion for
232
Bacteroidales (10.7%) was slightly higher than expected by chance (6.8%), whereas the
233
observed incidence of intra-order co-exclusion for Clostridiales (0.8%) was much lower than
234
expected by random associations (4.4%).
13
ACS Paragon Plus Environment
Environmental Science & Technology
Page 14 of 40
235
Discussion
236
Environmental parameters linked to methanogenic performance
237
In this study, four anaerobic digesters were operated for 90 days under different SRTs and
238
OLRs with the same inoculum and primary sewage sludge as biogas feedstock. Shortening
239
SRT to 9 days, which is below the putative minimum SRT, i.e., 10 days, to prevent washout of
240
slow-growing methanogens
241
abundance of HAc-utilizing methanogens, nor in the deterioration of biogas/methane
242
production. For example, relative abundances of Methanomicrobia and Methanosaeta (Figure
243
S4) and methane-producing rate (RM, Figure 1b) markedly increased with the decrease of SRT
244
from 16 or 12 days to 9 days. However, a further decrease of SRT from 9 to 7 days led to a
245
dramatic decrease of RM from 0.95 (±0.14) to 0.54 (±0.14) L/L/d (Table S1), despite a
246
further increase of the relative abundance of methanogens (Figure S4). This non-monotonic
247
relationship between SRT/OLR and methanogen abundance suggest that they are not the sole
248
influential factors of methanogens and RM. For example, the positive/negative correlations
249
between abundance of Bacteroidales populations (i.e., s1, s3 and s49, Figure 4) and RM/CH4%
250
suggest their important roles in mediating both methane-producing rate (as further discussed
251
below) and methane content in biogas. In line with this relationship, the increasing order of
252
the relative abundance of Bacteroidales populations by SRT during stage III (i.e., 7 < 16 < 12
253
< 9 days, Figure 6) well agrees with that of the biogas-producing rate between the digesters
254
(RM, Table S3). However, results obtained in this study are inadequate to suggest that all
6, 30
, may not necessarily result in the reduction of relative
14
ACS Paragon Plus Environment
Page 15 of 40
Environmental Science & Technology
255
digesters could reduce their SRT to below 10 days, considering the narrow range of possible
256
conditions and specific source of the seed and feed sludge. Therefore, the effects of SRT
257
should be assessed on a case-by-case basis.
258
The positive correlations between OLR and VFAs/HPr/HAc, as shown both by correlation
259
analysis (0.37-0.47, Table S4-1) and CCA (Figure 4), reflect increased hydrolysis and
260
conversion of organic macromolecules including carbohydrates, lipids and proteins to soluble
261
organic molecules with increasing OLR. In contrast, the negative correlations between RM and
262
VFAs/HPr/HAc (-0.39 to -0.48, Table S4-1) indicate that elevated levels of these soluble
263
intermediates may signify incomplete methanogenesis. The buildup of VFAs in R1 during
264
Stage I was probably an indicator of unbalanced growth between acidogenic bacteria and
265
methanogens. In contrast, no clear accumulation of VFAs in R2, R3 and R4 suggests
266
relatively stable conversion of VFAs to methane.
267
Community composition and dynamics linked to methanogenic performance
268
Microbial communities in the four digesters rapidly shifted away from the inoculated
269
community, followed the same trends in succession (along PC1, Figure 3), and converged into
270
highly similar community compositions (mainly Bacteroidia, Clostridia and Thermotogae)
271
under the selective pressures imposed by SRT, OLR and feed sludge (Figure S2 and Figure 2).
272
Both the nature of substrate (primary sludge) and SRT (7-16 days) were different from those
273
of the source digester of seed sludge that treats combined primary and secondary sludge at a
15
ACS Paragon Plus Environment
Environmental Science & Technology
Page 16 of 40
274
SRT of 6 days. Primary sludge has much higher biodegradability than secondary sludge since
275
it consists of more easily biodegradable organics, while SRT selects microorganisms by
276
growth rate and tends to wash out slowly growing organisms. Both our and other studies show
277
a positive relation between digester SRT and microbial richness 3, 4. Therefore, these selective
278
pressures may create specific niches that differentiate the community composition and
279
diversity from those of the inoculum.
280
First, both the microbial richness and evenness of 0.97-OTUs decreased over 90 days in all
281
digesters (Table 1). The decrease in diversity was mainly driven by the differences in the
282
nature of feeding sludge and SRT between this study and the source digester of the seed
283
sludge, reflecting the adaption of inoculated microorganisms to the regular feedings of CEPT
284
sludge. Moreover, many OTUs undetected or detected at low relative abundance in the seed
285
sludge (Figure 2) became the major trophic groups (Figure 6). Multiple populations capable
286
of the same metabolic functions, including hydrolyzing and fermenting organic molecules
287
(e.g., s1, s2, s3, s4, s7, s8, s13, s14, s15, s17, s22), and generating acetate (e.g., s5, s6, s12,
288
s16, s18 and s21) and methane (e.g., s30) were shared by the four digesters (Figure 2 and
289
Table S5), regardless of their different SRTs and OLRs, indicating high levels of functional
290
redundancy to maintain the process stability.
291
Hydrolysis and fermentation
292
Bacterial populations of orders Bacteroidales, Clostridiales and Thermotogales dominated
293
microbial communities in all digesters (Figure 6). Members of Clostridiales and 16
ACS Paragon Plus Environment
Page 17 of 40
Environmental Science & Technology
294
Thermotogales are well implicated in the hydrolysis and fermentation of carbohydrates in
295
anaerobic digesters
296
mammalian gut and anaerobic digesters as hydrolyzing/fermentative bacteria of carbohydrates
297
6, 32, 33
298
bacteria in anaerobic digesters (e.g., Clostridiales).
299
The rapid enrichment and continuous dominance of Bacteroidales populations (Figure 6) and
300
their significant (P-values < 0.01) negative Spearman’s correlations with cellulolytic
301
Clostridiales (-0.44) and Thermotogales (-0.64) populations reveals Bacteroidales spp. are
302
much stronger competitors of organic molecules in the anaerobic digesters than other bacteria.
303
Likewise, Bacteroidales populations are probably substrate competitors of Cloacamonales
304
(WWE1) and Saprospirales populations, considering the strong and significant negative
305
correlations (-0.88 and -0.80; P-values < 1×10-11) between them. The functions of WWE1 and
306
Saprospirales in methanogenic environments are poorly understood 34, 35. Our results suggest
307
that in anaerobic digesters they may share overlapped niches with Bacteroidales populations.
308
As operational conditions and resource availability in the anaerobic digesters favor the
309
prosperity and dominance of Bacteroidales, WWE1, Saprospirales and Thermotogales
310
populations are outcompeted and gradually washed out from the digesters (Figure 6).
311
On the contrary, Clostridiales spp. were dramatically enriched, secondary dominant
312
populations (Figure 6). The opposite intra-order species-spices co-exclusion (i.e., negative
313
associations between species of the same order) for Bacteroidales and Clostridiales
6, 31
. In contrast, Bacteroidales populations commonly occur in the
. However, we know little about their biological interactions with other cosmopolitan
17
ACS Paragon Plus Environment
Environmental Science & Technology
Page 18 of 40
314
implicates different ecological traits and rules guiding their co-dominance. Specifically,
315
Clostridiales populations have diverse metabolic capabilities of diverse organic molecules,
316
including cellulose and protein hydrolysis
317
and homoacetogenesis
318
niche differentiation may drive functionally versatile Clostridiales spp. into diverse patterns
319
of resource use to avoid exclusive competition. The metabolic versatility and phenotypic
320
plasticity with regard to the use of a broad range of organic micro- and macro-molecules in
321
Clostridiales populations should allow them to fulfill multiple functions and occupy broad
322
niches in anaerobic digesters. Based on these arguments, it is suggested that Clostridiales spp.
323
should be more ecologically divergent (or dissimilar) than Bacteroidales spp., thus
324
intra-species competition is stronger between members of Bacteroidales than Clostridiales.
325
Therefore, non-overlapped niches between Clostridiales and Bacteroidales populations could
326
make their co-dominance in the anaerobic digesters possible.
327
Acetogenesis and methanogenesis
328
Syntrophomonas, Syntrophobacterales and Clostridium can convert fatty acids to
329
methanogenic substrates (i.e., HAc and H2) 40, 41. These populations were the main acetogens
330
during anaerobic digestion (Figure 6), as also implicated by the positive correlations between
331
relative abundance of Syntrophomonas (s5, s12, s18 and s29) and Clostridium (e.g., s3, s4,
332
s19) OTUs and HAc concentration (Figure 4 and Table S5). The archaeal methanogens
333
mainly included acetoclastic Methanosarcinales and hydrogenotrophic Methanomicrobiales
36, 37
, polysaccharide fermentation 6, acetogenesis
38
, and syntrophic oxidation of fatty acids and acetate
39
. Therefore,
18
ACS Paragon Plus Environment
Page 19 of 40
Environmental Science & Technology
334
and Methanobacteriale, which are widely found together or individually as the core
335
methanogens in anaerobic digesters
336
between H2-producing Syntrophomonas populations and H2-scavenging Methanomicrobiales
337
and Methanobacteriales (0.45 and 0.60, P-values