1
Inoculum pretreatment differentially affects the
1
active microbial community performing
2
mesophilic and thermophilic dark fermentation
3
of xylose
4
Paolo Dessì a,*, Estefania Porca b, Luigi Frunzoc, Aino–Maija Lakaniemi a, Gavin Collins b, 5
Giovanni Espositod, Piet N. L. Lens a,e 6
7
aTampere University of Technology, Faculty of Natural Sciences, P.O. Box 541, FI-33101 8
Tampere, Finland 9
bMicrobial Communities Laboratory, School of Natural Sciences, National University of Ireland 10
Galway, University Road, Galway, H91 TK33, Ireland 11
cDepartment of Mathematics and Applications “Renato Caccioppoli”, University of Naples 12
Federico II, Via Cintia, Monte S. Angelo, Naples, Italy 13
dDepartment of Civil and Mechanical Engineering, University of Cassino and Southern Lazio, 14
Via di Biasio, 43, Cassino, FR, Italy 15
eUNESCO–IHE, Institute for Water Education, Westvest 7, 2611AX Delft, The Netherlands 16
17
Manuscript submitted to: Bioresource Technology 18
19
*Corresponding author:
20
Phone: +358 417239696, e-mail: paolo.dessi@tut.fi, mail: Tampere University of Technology, 21
P.O. Box 541, FI-33101 Tampere, Finland 22
2 Abstract
23
The influence of different inoculum pretreatments (pH and temperature shocks) on 24
mesophilic (37 °C) and thermophilic (55 °C) dark fermentative H2 production from 25
xylose (50 mM) and, for the first time, on the composition of the active microbial 26
community was evaluated. At 37 °C, an acidic shock (pH 3, 24 hours) resulted in the 27
highest yield of 0.8 mol H2 mol-1 xylose. The H2 and butyrate yield correlated with the 28
relative abundance of Clostridiaceae in the mesophilic active microbial community, 29
whereas Lactobacillaceae were the most abundant non-hydrogenic competitors 30
according to RNA-based analysis. At 55 °C, Clostridium and Thermoanaerobacter were 31
linked to H2 production, but only an alkaline shock (pH 10, 24 hours) repressed lactate 32
production, resulting in the highest yield of 1.2 mol H2 mol-1 xylose. This study showed 33
that pretreatments differentially affect the structure and productivity of the active 34
mesophilic and thermophilic microbial community developed from the same inoculum.
35 36
Keywords 37
Biohydrogen; pH shock; Temperature shock; Clostridium; Lactobacillus; MiSeq.
38 39 40 41 42 43 44 45 46
3 1. Introduction
47
The increasing energy demand results in the depletion of fossil fuel reserves and in the 48
emission of enormous quantities of greenhouse gases to the atmosphere. Due to its high 49
energy content and carbon neutrality, hydrogen (H2) is a promising green alternative to 50
fossil fuels [1]. Among H2 production technologies (for a review, see Nikolaidis and 51
Poullikkas, 2017), biological methods have the advantage of coupling H2 generation 52
with the treatment of organic carbon-containing wastes and wastewaters. Biological H2
53
production technologies include photo and dark fermentation, direct and indirect 54
photolysis, and the water gas-shift reaction [3]. The high H2 production rate, the simple 55
reactor technology (similar to the well-established anaerobic digestion), and the 56
abundance of microorganisms capable to produce H2 from a wide range of substrates 57
may promote the establishment of dark fermentation in industrial applications [4].
58 59
Promising H2 yields have been obtained through mesophilic (30–40 °C), thermophilic 60
(50–60 °C), and hyperthermophilic (> 65 °C) dark fermentation of sugars by pure 61
cultures or synthetic co-cultures of H2 producing microorganisms (for reviews, see Lee 62
et al., 2011 and Elsharnouby et al., 2013). However, a diverse microbial consortium is 63
often required for dark fermentation of more complex substrates, such as waste-derived 64
carbohydrate mixtures. Furthermore, the easier and more economic process control as 65
well as the lack of requirements for sterilisation make mixed cultures more suitable for 66
industrial waste treatment applications than pure cultures [7]. The main drawback of 67
mixed cultures and unsterile conditions is the possible development of microorganisms 68
competing with H2 producers for the substrates or even H2-consuming microorganisms 69
in the dark fermentation reactor, which can drastically reduce the overall H2 yield.
70
4 Homoacetogenic bacteria and hydrogenotrophic methane producing archaea are the two 71
most common H2 consuming microorganisms, but also propionate producers as well as 72
sulfate and nitrate reducers use H2 during their metabolism [3]. Also lactate producing 73
bacteria are commonly found in dark fermentation systems [8]. They also compete for 74
the substrate, and can even inhibit H2 producers by either causing acidification or 75
excreting bacteriocins [9].
76 77
Inoculum pretreatment technologies have been widely studied on mixed cultures to 78
select H2 producing microorganisms at the expense of H2 consumers. When exposed to 79
harsh conditions, some H2 producers, such as Clostridium, Bacillus and 80
Thermoanaerobacter, produce endospores as a defence mechanism [10,11]. This 81
increases their survival chances compared to non-spore-forming H2 consumers as they 82
are able to regerminate once the environmental conditions become favourable.
83
Temperature and pH shocks are the most widely applied pretreatment methods to enrich 84
the microbial community prior to dark fermentation [12]. High temperature disrupts the 85
cell wall of microorganisms, causing cell lysis and protein denaturation [13]. Low 86
temperature can cause protein gelling, intracellular formation of ice crystals and 87
membrane lipid stiffening [14]. Acidic and alkaline shocks affect the electric charge of 88
the cell membrane, may inactivate key enzymes, and may cause a change of the 89
intracellular pH, which could lead to cell wall disruption [15]. Other pretreatment 90
methods include substrate loading shock, chemical treatment (e.g. using 2- 91
bromoethansulphonate acid or chloroform, which inhibit methanogenic archaea), 92
aeration, electric shocks, ionising irradiation, microwaves, or ultrasonication (for a 93
review, see Wang and Yin 2017).
94
5 95
Many studies have compared the effect of the various pretreatments, and their 96
combination, on mesophilic [16–23] and thermophilic [24,25] dark fermentative H2
97
production. However, due to the different inoculum, substrate, and operating conditions, 98
many results appear controversial. Most of the studies compared the various 99
pretreatments in terms of H2 production, rather than focusing on the microbial 100
community. In a few cases, DNA-based analyses have been applied to reveal the 101
microbial community composition [16–18,22–24]. Microbial community analyses were 102
mainly conducted using low sensitive techniques, such as denaturing gradient gel 103
electrophoresis (DGGE), allowing detection of only the dominant species of the 104
community. Recent microbiological techniques, such as next-generation high- 105
throughput sequencing, likely allow more detailed information on how the microbial 106
community is affected by various pretreatments, in order to develop new strategies to 107
increase the H2 yields.
108 109
RNA-based approaches provide even more useful information than DNA-based studies 110
[26]. Indeed, they enable to depict the composition of the microbial community that 111
remains active after the various pretreatments, and thus to find out which species are 112
involved in H2 production and possible competitive pathways. Therefore, this study 113
pursues the double aim of (i) finding out how different pretreatments shape the active 114
microbial communities in dark fermentation reactors and (ii) how such microbial 115
communities evolve and produce H2 during mesophilic (37 °C) and thermophilic (55 116
°C) dark fermentation.
117 118
6 2. Materials and methods
119
2.1 Source and pretreatment of inoculum 120
Activated sludge was selected as the parent inoculum, as it contains a huge variety of 121
microbial species and enabled higher H2 production than digester sludge in a previous 122
study [27]. It was collected from a secondary settler of a municipal wastewater 123
treatment plant (Mutton Island, Galway, Ireland). It was dewatered by filtration through 124
a 0.1 mm diameter mesh. After filtration, total solids, volatile solids, and pH were 28.1 125
(± 2.4) g L-1, 22.7 (± 2.0) g L-1, and 6.9, respectively. The activated sludge was stored at 126
4 °C for approximately one week prior to being used in the experiment.
127 128
Temperature (heat treatment and freezing and thawing) and pH (acidic and alkaline) 129
shocks, widely used to select H2 producing organisms from several inocula [12], were 130
chosen among the possible pretreatment methods. The heat shock was conducted by 131
exposing the inoculum, collected in thin 15 mL tubes, to 90 °C for 15 minutes using a 132
pre-heated water bath (Clifton, UK). Freezing and thawing was done by incubating 133
sludge samples at -20 °C for 24 hours in thin 15 mL tubes and then defrosting them at 134
30 °C in a water bath. The acidic shock was given by adjusting the pH of the sludge to 135
3.0 with HCl, incubating at room temperature (about 20 °C) for 24 hours, and then 136
increasing the pH back to 7.0 with NaOH. The alkaline shock was done by adjusting the 137
pH to 10.0 with NaOH, incubating at room temperature for 24 hours, and then 138
decreasing the pH back to 7.0 with HCl. Both HCl and NaOH were used at a 139
concentration of 1 or 3 M. The sludge was continuously stirred by using a magnetic 140
stirrer while adjusting the pH.
141 142
7 2.2 Incubation and sampling
143
Glass serum bottles (120 mL volume) were filled with 45 mL of a modified DSMZ 144
medium n.144 (Dessì et al., 2017), and inoculated with 5 mL of pretreated activated 145
sludge. Anaerobic incubations with untreated sludge were also prepared as negative 146
controls. The initial pH was adjusted to 7.0 with 1 M NaOH, and bottles were flushed 147
with N2 for 5 min prior to incubation. Thermostatic conditions (37 or 55 °C) and 148
shaking at 60 rpm were provided by shaker incubators (Thermo Scientific, USA). The 149
batch cultures lasted until cumulative H2 production was not observed consecutively for 150
48 hours in any of the triplicate bottles (96-144 hours in total). All batch cultures were 151
conducted as triplicates.
152 153
Gas samples (5 mL) were collected daily, and stored into 5.9 mL EXETAINER®
154
hydrogen-tight gas sampling tubes (Labco, UK). Liquid samples were also collected 155
daily and stored at -20 °C in 1.5 mL Eppendorf tubes for analysis. Sludge samples were 156
collected before being inoculated into the serum bottles (0 hours samples) and from the 157
bottom of serum bottles at the end of the batch culture at 37 °C or 55 °C. They were 158
then stored at -80 °C in RNA-free 1.5 mL Eppendorf tubes untill microbiological 159
analysis.
160 161
2.3 Microbial community analyses based on 16S cDNA 162
Microbiological samples were pelleted by centrifugation (10,000xg, 10 min). Nucleic 163
acid co-extraction was done using a method modified from Griffiths et al. (2000). DNA 164
inhibition, cDNA synthesis, high-throughput sequencing (Illumina MiSeq) and 165
bioinformatics analysis of partial 16S rRNA gene sequences using primers 515f and 166
8 805r were performed as reported previously (Dessì et al., manuscript under revision), 167
but using a more recent version of Mothur (1.39.5) and the Silva database (v128). The 168
total number of sequences was 5,494,444 and they were reduced to 3,294,730 169
(2,276,495 unique sequences) after a quality check. The hierarchical clustering 170
dendrogram was done using the ”Pvclust” package of ”R” software [29].
171 172
2.4 Analytical methods 173
Gas produced was quantified by a syringe method (Owen et al., 1979) and its 174
composition was determined by gas chromatography as reported previously (Dessì et al.
175
2017). Cumulative H2 and CO2 production was calculated using a H2 mass balance 176
equation [30], and corrected to standard temperature (0 °C). Total solids and volatile 177
solids were determined by the standard methods (APHA, 1995). pH was measured 178
using a pH meter (WTW inoLab) equipped with a Slimtrode electrode (Hamilton). The 179
composition of the liquid phase (xylose, volatile fatty acids and alcohols) was 180
determined by high-performance liquid chromatograph (Shimadzu) equipped with a 181
Rezex TM ROA-Organic Acid H+ (8%) column (Phenomenex, USA), held at 70 °C, and 182
a refractive index detector (Shimadzu, Japan). The mobile phase and flow rate were 183
0.013 N H2SO4 and 0.6 mL min-1, respectively.
184 185
2.5 Statistical analysis 186
To assess significant differences in the effect of the tested pretreatments on H2 yield, 187
one-way analysis of variance (ANOVA) and the Tukey test [31] at p = 0.05 were 188
conducted using the IBM SPSS Statistics package. The output of the statistical analysis 189
is provided in the supporting material (File S1).
190
9 191
3. Results and discussion 192
3.1 Effect of the inoculum pretreatments on the active microbial community 193
No dominant microorganisms were detected in the untreated activated sludge, which 194
was characterised by a presence of various families of microorganisms with relative 195
abundance below 20% (Figure 1), mainly belonging to the phylum of Proteobacteria 196
and Bacteroidetes (Figure S1).
197 198
Despite the aerobic origin of the sludge, heat shock, freezing and thawing, and acidic 199
pretreatment favoured the establishment of families of spore forming anaerobic 200
microorganisms such as Clostridiaceae (Figure 1). Some non-spore-forming 201
microorganisms, such as Comamonadaceae sp. [32], were able to resist the heat shock 202
and the freezing and thawing treatment, with a relative abundance of 12-22% and 17- 203
19%, respectively, but their relative abundance was < 5% after the acidic or alkaline 204
shock (Figure 1).
205 206
After the alkaline shock, the active microbial community was very different than the 207
one obtained after the other pretreatments, as shown by the hierarchical clustering 208
dendrogram (Figure 2); it was dominated by Proteobacteria belonging to the families of 209
Moraxellaceae (relative abundance of 48-54%) and Aeromonadaceae (21-28%), and 210
more specifically by microorganisms closely related to Acinetobacter sp. and 211
Aeromonas sp. (Figure 1; Table 1). Although being non-spore forming [33,34], both 212
Acinetobacter and Aeromonas survived the alkaline pretreatment, but their relative 213
abundance dropped to < 1% after applying the other pretreatments. This suggests that 214
10 microorganisms are differently affected by the different pretreatments, and that also 215
some non-spore forming microorganisms could survive the thermal and pH stress 216
applied. The alkaline shock pretreatment repressed more microbial families than the 217
other pretreatments, as confirmed by the low diversity of the microbial community, 218
according to the diversity indexes (Table 2).
219 220
Figure 1, Figure 2, Table 1, Table 2 221
222
3.2 Effect of pretreatments on mesophilic (37 °C) dark fermentation of xylose 223
3.2.1 Effect of pretreatments on H2 yield at 37 °C 224
Based on H2 yields, the acidic shock was the most effective pretreatment for mesophilic 225
(37 °C) dark fermentation, resulting in a yield of 0.8 mol H2 mol-1 xylose (Figure 3). No 226
statistically significant differences were found between the alkaline and heat shock, 227
which resulted in a H2 yield of 0.5-0.6 mol H2 mol-1 xylose, and in both cases H2
228
production stopped within 48 hours. Freezing and thawing was the least effective 229
pretreatment method, resulting in a yield of only 0.2 mol H2 mol-1 xylose, similar to the 230
yield obtained with the untreated sludge (Figure 3).
231 232
Chang et al. (2011) compared various pretreatment methods (including acidic and 233
alkaline shock, heat shock, aeration, and chemical treatment) on waste activated sludge 234
and reported that an acidic shock (pH 3, 24 hours) was the most effective pretreatment, 235
yielding 1.5 mol H2 mol-1 glucose at 35 °C and at an initial pH of 7. Furthermore, 236
Chang et al. (2011) reported a 20% increase of the H2 yield from glucose after five 237
consecutive mesophilic batch cultures with acidic shock-pretreated activated sludge 238
11 (without repeating the acidic shock) suggesting that the pretreatment effect can be 239
sustained for extended time periods.
240 241
Figure 3 242
243
3.2.2 Effect of pretreatments on the xylose oxidation pathways at 37 °C 244
In all the mesophilic batch cultures, more than 90% of the xylose was consumed (Figure 245
4). Acetate was always produced, regardless of the pretreatment applied, with a final 246
concentration ranging between 10 and 21 mM. In dark fermentation, the acetate 247
pathway yields the highest amount of H2 (3.3 mol H2 mol-1 xylose). However, the H2
248
partial pressure may cause a shift of the dark fermentation pathway to butyrate 249
production [35], which yields only 1.67 mol H2 mol-1 xylose. This seems to be also the 250
case in this study, as butyrate was found in all the batch cultures (Figure 4) with a 251
concentration proportional to the H2 yield (Figure 5a). The H2 yield obtained from the 252
mesophilic batch cultures with the acidic shock-, alkaline shock-, and heat shock- 253
pretreated activated sludge is, however, too high to be entirely associated to the butyrate 254
pathway, which stoichiometrically yields 2 mol H2 mol-1 butyrate. This suggests that H2
255
was produced through both the acetate and the butyrate pathway. Acetate could be 256
produced also through non-hydrogenic pathways, such as homoacetogenesis, lowering 257
the net H2 production [3].
258 259
At 37 °C, ethanol was produced in all cultures regardless of the pretreatment, with a 260
maximum of 16 mM in the batch cultures with the heat shock-pretreated activated 261
sludge (Figure 4). Solventogenesis can occur as a detoxification process in case of low 262
12 pH (< 5) and accumulation of undissociated volatile fatty acids [36]. In fact, regardless 263
the pretreatment applied, the pH of the batch medium decreased to below 5.0 after 24- 264
48 hours incubation (Figure 4). The low pH was likely associated to lactate production.
265
In fact, the highest concentration of lactate of 14 and 10 mM was obtained in the batch 266
cultures with the untreated and alkaline shock-pretreated activated sludge, respectively 267
(Figure 4), resulting in the lowest observed pH of 4.0. Chaganti et al. (2012) as well 268
reported lactate production from glucose at 37 °C when using untreated and alkaline 269
shock-pretreated (pH 11, 24 hours) anaerobic sludge as inoculum.
270 271
Figure 4 272
273
3.2.3 Active microbial communities after batch culture at 37 °C 274
Cultivation at 37 °C resulted in a lower diversity of the active microbial community 275
from pretreated than untreated activated sludge, according to the diversity indexes 276
(Table 2). Methane was not detected in the mesophilic batch cultures, including the 277
untreated control, suggesting the absence or inactivity of methanogenic archaea. In fact, 278
the relative abundance of active archaea was < 1% in all the batch cultures. Yin et al.
279
(2014) and Chang et al. (2011) as well did not detect any methane in their mesophilic 280
batch cultures with the pretreated and untreated activated sludge. However, Ren et al.
281
(2008) reported methanogenic activity in mesophilic batch cultures with alkaline shock- 282
pretreated (pH 11, 24 hours) activated sludge.
283 284
Based on the data of this study, mesophilic H2 production correlated reversely with the 285
diversity of the active microbial community (Figure 5b). All mesophilic batch cultures 286
13 with pretreated activated sludge were dominated by microorganisms belonging to the 287
phylum Firmicutes, present with a relative abundance of 76-98% (Figure S1). Yin and 288
Wang (2016) pretreated anaerobically digested sewage sludge by gamma irradiation, 289
obtaining > 99% of Firmicutes in the microbial community after 36 hours of mesophilic 290
(33 °C) batch incubation with glucose (94.4 mM) as the substrate for dark fermentation.
291
This suggests that Firmicutes dominate the mesophilic dark fermentative microbial 292
communities, regardless of the pretreatment applied to the inoculum.
293 294
The family of Clostridiaceae was found regardless of the pretreatment applied, and 295
even in the batch cultures with the untreated activated sludge (Figure 1). The relative 296
abundance of Clostridiaceae had a positive correlation with both the H2 and butyrate 297
yield (Figure 5a). In fact, the acidic shock, which was the most effective pretreatment in 298
terms of H2 yield (Figure 3), promoted a high relative abundance of Clostridiaceae, 299
which accounted for 71-75% of the active microbial community after 144 hours 300
incubation at 37 °C (Figure 1). Clostridium sp. have been widely reported to dominate 301
mesophilic dark fermentative microbial communities, both in batch and continuous 302
reactors [27,38,39]. Liu et al. (2009) applied similar pretreatments to those applied in 303
this study to an intertidal marine sediment to perform mesophilic (37 °C) dark 304
fermentation of glucose. They reported similar results to this study in terms of 305
pretreatment efficacy and dominant microbial communities. However, their DNA-level 306
analysis, performed by PCR-DGGE, only allowed the detection of the dominant 307
species, whereas the RNA-level approach used in this study describes more accurately 308
the active microbial communities involved in the dark fermentative process.
309 310
14 Figure 5
311 312
Lactate producers of the Lactobacillaceae family were the main competitors to H2
313
producing microorganisms in all mesophilic batch cultures (Figure 1). Despite the 314
relatively high H2 yield, these microorganisms were found even in the batch cultures 315
with the acidic shock-pretreated activated sludge, with a relative abundance of 8-17%.
316
Kim et al. (2014) sequenced DNA from the microbial community developed upon 317
mesophilic (35 °C) dark fermentation of acidic shock-pretreated (pH 3, 12 hours) food 318
waste, without an external inoculum. They reported a relative abundance of 70% of 319
Clostridium and 20% Lactobacillus, similar to the relative abundances obtained in this 320
study. Furthermore, Kim et al. (2014) reported a higher relative abundance of 321
Clostridium (up to 90 %) and a higher H2 yield after performing an acidic shock at pH 2 322
and 1 for 12 hours, suggesting that the pretreatment conditions used in this study can be 323
further optimised to increase the H2 yield.
324 325
In contrast, an acidic shock (pH 3, 24 hours) of activated sludge has been also reported 326
to inhibit H2 producing microorganisms and favour the establishment of lactic acid 327
producing bacteria (Ren et al., 2008). In the batch cultures with the alkaline shock- 328
pretreated activated sludge, the low final pH of 4.0 (Figure 4) likely inhibited H2
329
producers and favoured lactic acid bacteria, resulting in the highest relative abundance 330
of Lactobacillaceae (53-61% of the relative abundance), which produced lactate at the 331
expenses of H2. Lactic acid bacteria may also inhibit Clostridium by excreting toxins 332
[9].
333 334
15 Freezing and thawing resulted in a high relative abundance of Proteobacteria (Figure 335
S1), such as Veillonellaceae (36-46%), which were found also in the batch cultures with 336
the untreated sludge (Figure 1). The representative sequence of Veillonellaceae matched 337
the genus Megamonas (Table 1), which has been previously reported to produce 338
propionate and acetate from the fermentation of glucose [41]. Both propionate (8 mM) 339
and acetate (10 mM) were indeed detected in the batch cultures with the freezing and 340
thawing-pretreated activated sludge (Figure 4). Both propionate production and the low 341
relative abundance (12 %) of Clostridium were likely the main cause for the low H2
342
yield obtained in these batch cultures (Figure 1).
343 344
3.3 Effect of pretreatments on thermophilic (55 °C) dark fermentation 345
3.3.1 Effect of pretreatments on H2 yield at 55 °C 346
At 55°C, the alkaline shock resulted in the highest yield of 1.2 mol H2 mol-1 xylose, two 347
times higher than the freezing and thawing pretreatment (Figure 3), which was similar 348
to the H2 yield obtained by the untreated activated sludge (0.5-0.6 mol H2 mol-1 xylose).
349
Both the acidic shock and heat shock resulted in the lowest H2 yield of 0.2-0.3 mol H2
350
mol-1 xylose. Interestingly, H2 was continuously produced for 96 hours in the 351
thermophilic batch cultures with the alkaline shock-pretreated activated sludge, whereas 352
it ceased in the first 24-48 hours in the other batch cultures. To our knowledge, only two 353
studies have been performed to compare the effect of pretreatments on thermophilic 354
dark fermentation [24,25], and none of them reported alkaline shock as the best 355
pretreatment. In the previous studies, granular anaerobic sludge [25] or anaerobic 356
digester sludge from a palm oil mill plant [24] was used as inoculum, whereas activated 357
sludge treating municipal wastewater was used in this study. O-Thong et al. (2009) 358
16 reported a higher thermophilic (60 °C) H2 yield from sucrose (1.96 mol H2 mol-1
359
hexose) with loading shock-pretreated (2 days incubation with 83.25 gCOD L-1 sucrose) 360
than acidic shock-, alkaline shock-, chemical treatment-, and heat shock-pretreated 361
anaerobic sludge.
362 363
Luo et al. (2010) obtained a higher thermophilic (60 °C) H2 yield from cassava stillage 364
by untreated than pretreated anaerobic granular sludge. This suggests that, depending on 365
the origin of the inoculum, thermophilic conditions can be sufficient to favour H2
366
producing bacteria over competitors. Luo et al. (2010) also claimed that the effect of 367
pretreatments was not significant in the long term, as the H2 yield was similar after 28 368
days of continuous dark fermentation of the cassava stillage regardless the pretreatment 369
applied.
370 371
It should be noted that, even if a low H2 yield is obtained in a batch culture, an efficient 372
H2 producing microbial community may develop in the long term. For example, heat 373
shock-pretreated activated sludge may switch its metabolism to the butyrate pathway 374
resulting in an increased H2 yield: Dessì et al. (2017) reported a 320% H2 yield increase 375
when performing four consecutive thermophilic (55 °C) batch cultures of heat shock- 376
pretreated (90 °C, 15 min) activated sludge. Therefore, to assess the potential of a 377
pretreated inoculum for dark fermentation, additional long-term experiments need to 378
accompany the initial screening test.
379 380
3.3.2 Effect of pretreatments on the xylose oxidation pathways at 55 °C 381
17 Thermophilic batch cultures with the untreated and alkaline shock- or freezing and 382
thawing-pretreated activated sludge resulted in a higher (> 90%) xylose consumption 383
than batch cultures with the acidic shock- (75%) and heat shock- (66%) pretreated 384
activated sludge (Figure 4). Interestingly, the alkaline shock was the only pretreatment 385
which suppressed lactate production, as most of the xylose was converted to acetate and 386
butyrate up to a final concentration of 6 and 13 mM, respectively (Figure 4).
387 388
In the thermophilic batch cultures with the alkaline shock-pretreated activated sludge, 389
the pH decreased at a lower rate than in the other batch cultures, resulting in the highest 390
final pH (4.4), which could have favoured H2 production. However, the H2 yield of 1.2 391
H2 mol-1 xylose was 30% higher than the maximum expected of 0.9 mol H2 mol-1 392
xylose, according to the acetate and butyrate concentrations produced (Figure 4). H2
393
overproduction has been previously reported in thermophilic dark fermentation [42], 394
and attributed to an unusual H2 production pathway, such as acetate oxidation.
395
However, in the batch cultures with alkaline-shock pretreated activated sludge, the 396
acetate concentration decreased from 9 to 6 mM only in the last 48 hours of incubation 397
(Figure 4), when the cumulative H2 yield was stable (Figure 3), suggesting that acetate 398
oxidation did not occur. Also Zheng et al. (2016) obtained an average of 14 % H2
399
overproduction from glucose by anaerobic sludge, but an isotope feeding assay with 400
13C-labeled acetate excluded the involvement of the acetate oxidation pathway.
401 402
In the thermophilic batch cultures with the freezing and thawing-pretreated inoculum, 403
lactate (8 mM) and ethanol (10 mM) were produced together with acetate (12 mM) and 404
butyrate (8 mM), suggesting that none of the metabolic pathways was prevailing over 405
18 the other. In the batch cultures with untreated activated sludge, xylose was initially 406
fermented to acetate and butyrate, which both reached a concentration of 5-6 mM after 407
24 hours incubation. Then, a switch from the butyrate to the lactate pathway occurred, 408
and lactate was produced together with acetate in the subsequent 48 hours, resulting in a 409
final concentration of 15 mM lactate and acetate (Figure 4). In the batch cultures with 410
the acidic shock- and heat shock-pretreated sludge, lactate was the main metabolite, 411
with a concentration of 8 and 9 mM, respectively (Figure 4), resulting in the lowest H2
412
yield (Figure 3). Lactate production resulted in a fast pH decrease: the final pH was 4.2- 413
4.3 in the batch cultures with the acidic shock-, heat shock- and freezing and thawing- 414
pretreated activated sludge, and even lower (3.9) in the batch cultures with the untreated 415
activated sludge (Figure 4), in which lactate production was the highest (Figure 4). Such 416
a low pH likely inhibited thermophilic H2 production [44].
417 418
3.3.3 Active microbial communities after batch culture at 55 °C 419
After incubation at 55 °C, regardless of the pretreatment applied, > 98% of the active 420
microbial community was composed of microorganisms of the phylum Firmicutes 421
(Figure S1). In general, the thermophilic active microbial communities were less diverse 422
than the mesophilic ones (Table 2). The active microbial communities from 423
thermophilic batch cultures with the untreated and acidic shock-, alkaline shock-, and 424
freezing and thawing-pretreated activated sludge, were all similar, and dominated by the 425
families of Clostridiaceae and Peptostreptococcaceae (Figure 2).
426 427
A different active microbial community developed in the thermophilic batch culture 428
with the heat shock-pretreated activated sludge (Figure 1), in which a family of 429
19 Thermoanaerobacterales (Figure 1), closely related to T. thermosaccharolyticum (Table 430
1), was prevailing with a relative abundance up to 94%. Similarly, O-Thong et al.
431
(2009) compared various pretreatments on thermophilic (60 °C) dark fermentation of 432
sucrose in batch experiments, and reported the predominance of T.
433
thermosaccharolyticum from heat shock-pretreated anaerobic sludge, whereas 434
Clostridium became dominant from both acidic and alkaline shock-pretreated anaerobic 435
sludge. The family of Clostridiaceae includes thermophilic species, such as C.
436
thermosaccharolyticum [45] and C. thermopalmarium [46], able to convert sugars to 437
H2, acetate and butyrate. T. thermosaccharolyticum effectively produces H2 from xylose 438
at 55-60 °C and in the pH range 5.5-7.0 (Ren et al., 2008). In the batch cultures with the 439
heat shock-pretreated activated sludge of this study, T. thermosaccharolyticum was thus 440
likely inhibited by the low pH, which dropped to < 4.5 within 48 hours (Figure 4), 441
resulting in a low H2 yield (Figure 3) and poor xylose consumption (Figure 4).
442 443
Bacillales, found in the thermophilic batch cultures with the heat shock- (3-30%) and 444
acidic shock- (3-11%) pretreated activated sludge (Figure 1), might be responsible for 445
lactate production. In fact, the order of Bacillales includes lactic acid producers such as 446
Sporolactobacillus, a spore-forming acidophilic microorganism previously reported in 447
thermophilic dark fermentation of sugars [48].
448 449
4. Conclusions 450
H2 production at 37 °C was dependent on the relative abundance of Clostridiaceae in 451
the microbial community, and an acidic shock favoured their establishment at the 452
expense of competing microorganisms. At 55 °C, an alkaline shock suppressed lactate 453
20 production, resulting in a high H2 yield, and is thus a suitable pretreatment for starting- 454
up thermophilic H2 producing bioreactors. Inoculum pretreatment appears necessary for 455
mesophilic dark fermentation, whereas less invasive strategies (such as pH control at a 456
pH of 5 or 6) could be considered to select H2 producing microorganisms at the expense 457
of lactic acid bacteria in thermophilic bioreactors.
458 459
Acknowledgements 460
The authors gratefully thank Leena Ojanen (Tampere University of Technology, 461
Finland) for analysing the gas samples, Camilla Thorn (National University of Ireland 462
Galway, Ireland) for helping with the microbial community analysis, and the Mutton 463
island wastewater treatment plant (Galway, Ireland) for providing the activated sludge 464
inoculum.
465 466
Funding 467
This work was supported by the Marie Skłodowska-Curie European Joint Doctorate 468
(EJD) in Advanced Biological Waste-To-Energy Technologies (ABWET) funded from 469
Horizon 2020 under grant agreement no. 643071.
470 471
Declaration of interest 472
The authors declare no conflicts of interest.
473 474
References 475
[1] Palomo-Briones R, Razo-Flores E, Bernet N, Trably E. Dark-fermentative 476
biohydrogen pathways and microbial networks in continuous stirred tank 477
21 reactors: Novel insights on their control. Appl Energy 2017;198:77–87.
478
[2] Nikolaidis P, Poullikkas A. A comparative overview of hydrogen production 479
processes. Renew Sustain Energy Rev 2017;67:597–611.
480
[3] Bundhoo MAZ, Mohee R. Inhibition of dark fermentative bio-hydrogen 481
production: A review. Int J Hydrogen Energy 2016;41:6713–33.
482
[4] Arimi MM, Knodel J, Kiprop A, Namango SS, Zhang Y, Geißen S-U. Strategies 483
for improvement of biohydrogen production from organic-rich wastewater: A 484
review. Biomass and Bioenergy 2015;75:101–18.
485
[5] Lee DJ, Show KY, Su A. Dark fermentation on biohydrogen production: Pure 486
culture. Bioresour Technol 2011;102:8393–402.
487
[6] Elsharnouby O, Hafez H, Nakhla G, El Naggar MH. A critical literature review 488
on biohydrogen production by pure cultures. Int J Hydrogen Energy 489
2013;38:4945–66.
490
[7] Li C, Fang HHP. Fermentative Hydrogen Production From Wastewater and Solid 491
Wastes by Mixed Cultures. Crit Rev Environ Sci Technol 2007;37:1–39.
492
[8] Etchebehere C, Castelló E, Wenzel J, del Pilar Anzola-Rojas M, Borzacconi L, 493
Buitrón G, et al. Microbial communities from 20 different hydrogen-producing 494
reactors studied by 454 pyrosequencing. Appl Microbiol Biotechnol 495
2016;100:3371–84.
496
[9] Noike T, Takabatake H, Mizuno O, Ohba M. Inhibition of hydrogen fermentation 497
of organic wastes by lactic acid bacteria. Int J Hydrogen Energy 2002;27:1367–
498
71.
499
[10] Collins MD, Lawson PA, Willems A, Cordoba JJ, Fernandez-Garayzabal J, 500
Garcia P, et al. The Phylogeny of the Genus Clostridium: Proposal of Five New 501
22 Genera and Eleven New Species Combinations. Int J Syst Bacteriol
502
1994;44:812–26.
503
[11] Galperin MY. Genome Diversity of Spore-Forming Firmicutes. Microbiol Spectr 504
2013;1:1–27. doi:10.1128/microbiolspectrum.TBS-0015-2012.Genome.
505
[12] Wang J, Yin Y. Principle and application of different pretreatment methods for 506
enriching hydrogen-producing bacteria from mixed cultures. Int J Hydrogen 507
Energy 2017;42:4804–23.
508
[13] Appels L, Baeyens J, Degrève J, Dewil R. Principles and potential of the 509
anaerobic digestion of waste-activated sludge. Prog Energy Combust Sci 510
2008;34:755–81.
511
[14] Sawicka JE, Robador A, Hubert C, Jørgensen BB, Brüchert V. Effects of freeze–
512
thaw cycles on anaerobic microbial processes in an Arctic intertidal mud flat.
513
ISME J 2010;4:585–94.
514
[15] Rafieenia R, Lavagnolo MC, Pivato A. Pre-treatment technologies for dark 515
fermentative hydrogen production: Current advances and future directions. Waste 516
Manag 2017. doi:10.1016/j.wasman.2017.05.024.
517
[16] Misturini Rossi D, Berne da Costa J, Aquino de Souza E, Ruaro Peralba M do C, 518
Samios D, Záchia Ayub MA. Comparison of different pretreatment methods for 519
hydrogen production using environmental microbial consortia on residual 520
glycerol from biodiesel. Int J Hydrogen Energy 2011;36:4814–9.
521
[17] Guo L, Lu M, Li Q, Zhang J, She Z. A comparison of different pretreatments on 522
hydrogen fermentation from waste sludge by fluorescence excitation-emission 523
matrix with regional integration analysis. Int J Hydrogen Energy 2015;40:197–
524
208.
525
23 [18] Jeong DY, Cho SK, Shin HS, Jung KW. Application of an electric field for 526
pretreatment of a seeding source for dark fermentative hydrogen production.
527
Bioresour Technol 2013;139:393–6.
528
[19] Chang S, Li J-Z, Liu F. Evaluation of different pretreatment methods for 529
preparing hydrogen-producing seed inocula from waste activated sludge. Renew 530
Energy 2011;36:1517–22.
531
[20] Chaganti SR, Kim D-H, Lalman JA. Dark fermentative hydrogen production by 532
mixed anaerobic cultures: Effect of inoculum treatment methods on hydrogen 533
yield. Renew Energy 2012;48:117–21.
534
[21] Yin Y, Hu J, Wang J. Enriching hydrogen-producing bacteria from digested 535
sludge by different pretreatment methods. Int J Hydrogen Energy 536
2014;39:13550–6.
537
[22] Liu H, Wang G, Zhu D, Pan G. Enrichment of the hydrogen-producing microbial 538
community from marine intertidal sludge by different pretreatment methods. Int J 539
Hydrogen Energy 2009;34:9696–701.
540
[23] Ren N, Guo W-Q, Wang X-J, Xiang W-S, Liu B-F, Wang X-Z, et al. Effects of 541
different pretreatment methods on fermentation types and dominant bacteria for 542
hydrogen production. Int J Hydrogen Energy 2008;33:4318–24.
543
[24] O-Thong S, Prasertsan P, Birkeland N. Evaluation of methods for preparing 544
hydrogen-producing seed inocula under thermophilic condition by process 545
performance and microbial community analysis. Bioresour Technol 546
2009;100:909–18.
547
[25] Luo G, Xie L, Zou Z, Wang W, Zhou Q. Evaluation of pretreatment methods on 548
mixed inoculum for both batch and continuous thermophilic biohydrogen 549
24 production from cassava stillage. Bioresour Technol 2010;101:959–64.
550
[26] De Vrieze J, Regueiro L, Props R, Vilchez-Vargas R, Jáuregui R, Pieper DH, et 551
al. Presence does not imply activity: DNA and RNA patterns differ in response to 552
salt perturbation in anaerobic digestion. Biotechnol Biofuels 2016;9:244.
553
[27] Dessì P, Lakaniemi A-M, Lens PNL. Biohydrogen production from xylose by 554
fresh and digested activated sludge at 37, 55 and 70 °C. Water Res 555
2017;115:120–9.
556
[28] Griffiths RI, Whiteley AS, O’Donnell AG, Bailey MJ. Rapid Method for 557
Coextraction of DNA and RNA from Natural Environments for Analysis of 558
Ribosomal DNA- and rRNA-Based Microbial Community Composition. Appl 559
Environ Microbiol 2000;66:5488–91.
560
[29] Suzuki R, Shimodaira H. Pvclust: an R package for assessing the uncertainty in 561
hierarchical clustering. Bioinformatics 2006;22:1540–2.
562
[30] Logan BE, Oh SE, Kim IS, Van Ginkel S. Biological Hydrogen Production 563
Measured in Batch Anaerobic Respirometers. Environ Sci Technol 564
2002;36:2530–5.
565
[31] Box GEP, Holohan R, Scott E. Statistics for experimenters: an introduction to 566
design, data analysis, and model building. John Wiley and sons; 1978.
567
[32] Willems A, De Ley J, Gillis M, Kersters K. Comamonadaceae , a new family 568
encompassing the Acidovorans rRNA complex, including Variovorax paradoxus 569
gen. nov., comb. nov., for Alcaligenes paradoxus (Davis 1969). Int J Syst Evol 570
Microbiol 1991;41:445–50.
571
[33] Bouvet PJM, Grimont PAD. Taxonomy of the genus Acinetobacter with the 572
recognition of Acinetobacter baumannii sp. nov., Acinetobacter haemolyticus sp.
573
25 nov., Acinetobacter johnsonii sp. nov., and Acinetobacter junii sp. nov. and 574
emended descriptio. Int J Syst Bacteriol 1986;36:228–40.
575
[34] Alperi A, Martínez-Murcia AJ, Ko W-C, Monera A, Saavedra MJ, Figueras MJ.
576
Aeromonas taiwanensis sp. nov. and Aeromonas sanarellii sp. nov., clinical 577
species from Taiwan. Int J Syst Evol Microbiol 2010;60:2048–55.
578
[35] Angenent LT, Karim K, Al-dahhan MH, Wrenn BA, Domı R. Production of 579
bioenergy and biochemicals from industrial and agricultural wastewater. Trends 580
Biotechnol 2004;22:477–85.
581
[36] Van Ginkel S, Logan BE. Inhibition of Biohydrogen production by 582
Undissociated Acetic and Butyric Acids. Environ Sci Technol 2005;39:9351–6.
583
[37] Yin Y, Wang J. Changes in microbial community during biohydrogen production 584
using gamma irradiated sludge as inoculum. Bioresour Technol 2016;200:217–
585
22.
586
[38] Si B, Liu Z, Zhang Y, Li J, Xing X-H, Li B, et al. Effect of reaction mode on 587
biohydrogen production and its microbial diversity. Int J Hydrogen Energy 588
2015;40:3191–200.
589
[39] Sivagurunathan P, Kumar G, Park J-H, Park J-H, Park H-D, Yoon J-J, et al.
590
Feasibility of enriched mixed cultures obtained by repeated batch transfer in 591
continuous hydrogen fermentation. Int J Hydrogen Energy 2016;41:4393–403.
592
[40] Kim D-H, Jang S, Yun Y-M, Lee M-K, Moon C, Kang W-S, et al. Effect of acid- 593
pretreatment on hydrogen fermentation of food waste: Microbial community 594
analysis by next generation sequencing. Int J Hydrogen Energy 2014;39:16302–
595 596 9.
[41] Chevrot R, Carlotti A, Sopena V, Marchand P, Rosenfeld E. Megamonas 597
26 rupellensis sp. nov., an anaerobe isolated from the caecum of a duck. Int J Syst 598
Evol Microbiol 2008;58:2921–4.
599
[42] Kotsopoulos TA, Zeng RJ, Angelidaki I. Biohydrogen production in granular up- 600
flow anaerobic sludge blanket (UASB) reactors with mixed cultures under hyper- 601
thermophilic temperature (70°C). Biotechnol Bioeng 2006;94:296–302.
602
[43] Zheng H, Zeng RJ, O’Sullivan C, Clarke WP. Critical analysis of hydrogen 603
production from mixed culture fermentation under thermophilic condition (60 604
°C). Appl Microbiol Biotechnol 2016;100:5165–76.
605
[44] Zhang T, Liu H, Fang HHP. Biohydrogen production from starch in wastewater 606
under thermophilic condition. J Environ Manage 2003;69:149–56.
607
[45] Islam S, Zhang C, Sui K, Guo C, Liu C. Coproduction of hydrogen and volatile 608
fatty acid via thermophilic fermentation of sweet sorghum stalk from co-culture 609
of Clostridium thermocellum and Clostridium thermosaccharolyticum. Int J 610
Hydrogen Energy 2016;42:830–7.
611
[46] Lawson Anani Soh A, Ralambotiana H, Ollivier B, Prensier G, Tine E, Garcia J- 612
L. Clostridium thermopalmarium sp. nov., a Moderately Thermophilic Butyrate- 613
Producing Bacterium Isolated from Palm Wine in Senegal. Syst Appl Microbiol 614
1991;14:135–9.
615
[47] Ren N, Cao G, Wang A, Lee D, Guo W, Zhu Y. Dark fermentation of xylose and 616
glucose mix using isolated Thermoanaerobacterium thermosaccharolyticum 617
W16. Int J Hydrogen Energy 2008;33:6124–32.
618
[48] Abreu AA, Karakashev D, Angelidaki I, Sousa DZ, Alves MM. Biohydrogen 619
production from arabinose and glucose using extreme thermophilic anaerobic 620
mixed cultures. Biotechnol Biofuels 2012;5:6.
621
27 Table and Figure captions
622
Table 1. Association of a representative 16S rRNA gene sequence of the most abundant 623
12 microbial families obtained in this study (Figure 1) to those collected in the 624
GenBank.
625 626
Table 2. Diversity of the active microbial community (cDNA) after various 627
pretreatments, and after incubation at 37 or 55 °C, measured by the Shannon, Simpson 628
and Pielou’s J’ index. The number in parenthesis is the standard deviation of the 629
triplicate.
630 631
Figure 1. Relative abundance of the active families resulting from MiSeq sequencing of 632
the partial 16S rRNA on microbiological samples collected before and after batch 633
cultures with the untreated and pretreated sludge at 37 and 55 °C. The results are 634
reported in triplicate. “Other” refers to the sum of genera with relative abundance < 1%.
635 636
Figure 2. Hierarchical clustering of active microbial communities from untreated (U) 637
and pretreated (AS, acidic shock; BS, alkaline shock; HS, heat shock; FT, freezing and 638
thawing) inoculum before (0 h) or after mesophilic (37 °C) or thermophilic (55 °C) dark 639
fermentation of xylose. The values reported are the approximately unbiased (au) p- 640
value, bootstrap probability (bp), and cluster label (edge #). Clusters with au > 95% are 641
indicated in the rectangles. The analysis was conducted on the average relative 642
abundance of triplicates.
643 644
28 Figure 3. H2 yield from mesophilic (37 °C) and thermophilic (55 °C) dark fermentation 645
of xylose by the untreated or pretreated activated sludge. The results shown are the 646
average of three independent batch cultures. The error bars represent the standard 647
deviation of the triplicates.
648 649
Figure 4. Xylose, metabolite concentration and pH profiles from mesophilic (37 °C) 650
and thermophilic (55 °C) dark fermentation of xylose by the untreated or pretreated 651
activated sludge. The results shown are the average of three independent batch cultures.
652
The error bars represent the standard deviation of the triplicates.
653 654
Figure 5. Correlation between the diversity index (Simpson) and the H2 yield (a), and 655
between the relative abundance of active Clostridiaceae and H2 or butyrate yield (b) 656
after mesophilic (37 °C) dark fermentation of xylose by the untreated (C) and pretreated 657
(AS, acidic shock; BS, alkaline shock; HT, heat treatment; FT, freezing and thawing) 658
inoculum.
659 660
29 Table 1.
661
Familya Genus and speciesb Accession
number
Matching sequencec
Similarity (%)d
Clostridiaceae Clostridium sp. HF566199 458-749 99
Lactobacillaceae Lactobacillus mucosae MF425117 490-781 99 Peptostreptococcaceae Romboutsia ilealis LN555523 2071187-
2071477
99
Thermoanaerobacterales Family III
Thermoanaerobacterium thermosaccharolyticum
MF405082 414-705 99
Moraxellaceae Acinetobacter sp. KY818302 465-756 99
Comamonadaceae Acidovorax sp. LC279183 425-716 99
Veillonellaceaee Megamonas sp. LT628480 522-812 99
Acidimicrobiales Incertae Sedis
Candidatus Microthrix parvicella
KM052469 91-383 99
Rhodocyclaceaef Dechloromonas KY029047 486-777 99
Myxococcales unclassifiedg
- - - -
Aeromonadaceae Aeromonas MF461171 478-769 100
Chitinophagaceae Terrimonas arctica NR_134213 454-745 99
a The families refer to Figure 4 in the main text; b Closest cultured species in GenBank; c 662
Position (in bp) in which the sequence overlaps the reference sequence; d Percentage of identical 663
nucleotide pairs between the 16S rRNA gene sequence and the closest cultured species in 664
GenBank; e Classified as Selenomonadaceae in the GenBank; e Classified as Azonexaceae in the 665
GenBank; f No match at family, genus and species level.
666 667 668
30 Table 2.
669
Samplea Pretreatment Shannon
diversity
Simpson diversity
J’ Evenness
0 hours (inoculum)
Untreated 3.10 (± 0.06) 0.92 (± 0.01) 0.73 (± 0.02) Acidic shock 3.08 (± 0.02) 0.92 (± 0.00) 0.72 (± 0.01) Alkaline shock 1.56 (± 0.07) 0.66 (± 0.02) 0.41 (± 0.01) Heat shock 2.86 (± 0.02) 0.88 (± 0.00) 0.66 (± 0.01) Freezing and thawing 3.36 (± 0.03) 0.94 (± 0.00) 0.78 (± 0.01) Batch culture
37 °C
Untreated 2.81 (± 0.07) 0.90 (± 0.01) 0.66 (± 0.02) Acidic shock 1.01 (± 0.10) 0.45 (± 0.03) 0.30 (± 0.03) Alkaline shock 1.21 (± 0.21) 0.56 (± 0.06) 0.35 (± 0.04) Heat shock 1.30 (± 0.07) 0.64 (± 0.02) 0.46 (± 0.04) Freezing and thawing 2.02 (± 0.11) 0.77 (± 0.03) 0.50 (± 0.02) Batch culture
55 °C
Untreated 1.12 (± 0.08) 0.56 (± 0.06) 0.30 (± 0.02) Acidic shock 1.14 (± 0.11) 0.58 (± 0.06) 0.38 (± 0.02) Alkaline shock 1.10 (± 0.20) 0.56 (± 0.08) 0.34 (± 0.02) Heat shock 0.68 (± 0.53) 0.31 (± 0.29) 0.23 (± 0.16) Freezing and thawing 1.01 (0.08) 0.54 (± 0.06) 0.30 (± 0.03)
a All samples were subsampled to the size of the smallest sample (8,654 sequences) 670
671 672 673 674 675 676 677
31 Figure 1
678
679 680
32 Figure 2
681
682 683 684 685 686 687 688 689 690 691 692 693
33 Figure 3
694
695 696 697 698
34 Figure 4
699
700
35 Figure 5
701
702
