Thiosulfate Conversion to Sulfide by a ... - ACS Publications

Dec 20, 2016 - Wetsus, European Centre of Excellence for Sustainable Water ... microbial community was dominated by bacteria belonging to the family...
0 downloads 0 Views 2MB Size
Subscriber access provided by GAZI UNIV

Article

Thiosulfate conversion to sulfide by a haloalkaliphilic microbial community in a bioreactor fed with H gas 2

João A. B. Sousa, Martijn F. M. Bijmans, Alfons J.M. Stams, and Caroline M. Plugge Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b04497 • Publication Date (Web): 20 Dec 2016 Downloaded from http://pubs.acs.org on December 21, 2016

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

1 2

Thiosulfate conversion to sulfide by a haloalkaliphilic microbial community in a bioreactor fed with H 2 gas

3 4

João A.B. Sousa*1,2, Martijn F.M. Bijmans2, Alfons J.M. Stams1,3 and Caroline M. Plugge1,2

5 6

1

7

The Netherlands

8

2

9

8911 MA Leeuwarden, The Netherlands

Laboratory of Microbiology, Wageningen University, Stippeneng 4, 6708 WE Wageningen,

Wetsus, European Centre of excellence for Sustainable Water Technology, Oostergoweg 9,

10

3

11

Braga, Portugal

12

(E-mail: [email protected]; Phone: +31611799992 )

CEB-Centre of Biological Engineering, University of Minho, Campus de Gualtar, 4710-057,

13 14

1 ACS Paragon Plus Environment

Environmental Science & Technology

Page 2 of 40

15

Abstract

16

In industrial gas biodesulfurization systems, where haloalkaline conditions prevail, a

17

thiosulfate containing bleed stream is produced. This bleed stream can be treated in a separate

18

bioreactor by reducing thiosulfate to sulfide and recycling it. By performing treatment and

19

recycling of the bleed stream, its disposal decreases and less caustics is required to maintain

20

the high pH. In this study, anaerobic microbial thiosulfate conversion to sulfide in a H2/CO2

21

fed bioreactor operated at haloalkaline conditions was investigated. Thiosulfate was converted

22

by reduction to sulfide as well as disproportionation to sulfide and sulfate. Formate

23

production from H2/CO2 was observed as an important reaction in the bioreactor. Formate,

24

rather than H2, might have been used as the main electron donor by thiosulfate/sulfate-

25

reducing bacteria. The microbial community was dominated by bacteria belonging to the

26

family

27

phylogenetically related to known haloalkaline sulfate and thiosulfate reducers, thiosulfate-

28

disproportionating bacteria and remarkably sulfur-oxidizing bacteria were also detected.

29

Based on the results, two approaches to treat the biodesulfurization waste stream are

30

proposed: i) addition of electron donor to reduce thiosulfate to sulfide and ii) thiosulfate

31

disproportionation without the need for an electron donor. The concept of application of

32

solely thiosulfate disproportionation is discussed.

Clostridiaceae

most

closely

related

to

Tindallia

texcoconensis.

Bacteria

33

34

Keywords

35

Thiosulfate reduction, Thiosulfate disproportionation, Sulfate-reducing bacteria, Tindallia,

36

Haloalkaline, Formate

2 ACS Paragon Plus Environment

Page 3 of 40

Environmental Science & Technology

37

3 ACS Paragon Plus Environment

Environmental Science & Technology

Page 4 of 40

38

Introduction

39

Thiosulfate (S2O32-) is one of the major reactive intermediate in the biogeochemical sulfur

40

cycle (1). Thiosulfate can be formed abiotically through oxidation of sulfide (HS-),

41

polysulfide (Sx2-) or pyrite with O2. Most sulfate-reducing bacteria (SRB) are able to reduce

42

thiosulfate to sulfide with H2 or an organic electron donor (Eq. 1) (2). However, thiosulfate

43

can also be disproportionated to sulfate and sulfide in the absence of an electron donor (Eq.

44

2) (3). Disproportionation of thiosulfate can also be performed by several SRB (4, 5).

45 46

    +   →   +   

(Eq. 1)

47

     +   →  +  +

(Eq. 2)

48 49

Thiosulfate is particularly relevant in the sulfur cycle of haloalkaline environments, such as

50

soda lakes where sodium carbonate and bicarbonate are the dominant salts. In these

51

environments, thiosulfate is abiotically formed by the oxidation of polysulfides that are

52

chemically stable at high pH and anoxic conditions (6). In soda lakes thiosulfate reduction is

53

even more prominent than sulfate reduction (7, 8). Several anaerobic thiosulfate-converting

54

microorganisms have been isolated from soda lakes (8). These isolated microbes, mainly

55

belonging to genus Desulfonatronun, Desulfonatronovibrio and Desulfonatronospira, are able

56

to perform both thiosulfate disproportionation and thiosulfate reduction.

57

Thiosulfate is also an important intermediate in biodesulfurization processes that operate at

58

haloalkaline conditions (9). In these processes thiosulfate is an unwanted soluble product

59

along with sulfate and needs to be removed via a bleed stream. This bleed stream is rich in

4 ACS Paragon Plus Environment

Page 5 of 40

Environmental Science & Technology

60

thiosulfate which contributes to the chemical oxygen demand (COD) of the disposed water.

61

Additionally, the production of a bleed stream increases the demand in water and caustic to

62

maintain the pH high enough to absorb hydrogen sulfide from the gas. The increased demand

63

of water and caustics is critical when biodesulfurization plants are built in locations without

64

easy access to water and caustics, requiring transport of both.

65

Thiosulfate in the bleed stream can be removed by biological conversion to sulfide in a

66

separate anaerobic bioreactor. The sulfide produced in this anaerobic bioreactor can then be

67

recycled back to the sulfide oxidation bioreactor. In this bioreactor, the oxygen supply is

68

controlled to maximize production of elemental sulfur and minimize sulfate and thiosulfate

69

formation. The elemental biosulfur is recovered and used in agriculture or in chemical and

70

biotechnological industries (10).

71

At present, there are no studies available on long-term cultivation of thiosulfate-reducing

72

biomass in reactor systems at haloalkaline conditions, which is needed for full-scale

73

application of the proposed concept Detailed information on the microbial activity and

74

composition, metabolic interactions of microorganisms and biomass characteristics is lacking.

75

This information is crucial to understand and improve thiosulfate conversion in bioreactors.

76

In this work microbial thiosulfate conversion at haloalkaline conditions in a continuous

77

reactor system was investigated. A gas-lift bioreactor fed with H2 as electron donor and

78

thiosulfate as electron acceptor was used. The bioreactor was operated at pH 9 and 1.5 M Na+

79

and no organic compounds were present in the feed. Thiosulfate conversion was monitored

80

and biomass growth and biomass characteristics were determined. The dominant microbes

81

were identified using molecular biological techniques.

82

5 ACS Paragon Plus Environment

Environmental Science & Technology

Page 6 of 40

83

6 ACS Paragon Plus Environment

Page 7 of 40

84

85

Environmental Science & Technology

Experimental Bioreactor set-up

86

A 4.4 l glass gas lift reactor with an internal 3 phase-separator was used (Figure S1).

87

Temperature was maintained at 35oC using a water jacket connected to a thermostat bath

88

(DC10-P5/U, Haake, Dreieich, Germany). The influent feeding was performed by a

89

membrane pump (Stepdos 08 RC, KNF-Verder, Utrecht, the Netherlands). The H2 and CO2

90

gas supply was controlled using digital mass flow controllers (F-201CV-020-AGD-22-V and

91

F-201CV-020-AGD-22-Z, Bronkhorst, Ruurlo, the Netherlands). The gas was recycled using

92

a vacuum pump (Laboport®, KNF, Trenton, NJ) and the gas flow was measured with a

93

calibrated flow meter (URM, Kobold, Arnhem, the Netherlands). A pH and a redox potential

94

sensor (CPS11D and CPS12D, Endress+Hauser, Naarden, the Netherlands) connected to a

95

controller (Liquiline CM44x, Endress+Hauser, Naarden, the Netherlands) were used to

96

monitor the conditions inside the reactor. The pH was controlled at pH 9 by supplying CO2

97

using the mass flow controller with a supply rate range between 0.05 - 5 ml min-1.

98

The bioreactor was inoculated with 50 ml of biomass collected from a gas lift bioreactor

99

earlier operated with sulfate and thiosulfate (1/1 mol ratio) and fed with H2 and CO2 (11).

100

Medium

101

The mineral medium used was buffered at pH 9 (± 0.05) with sodium carbonate and sodium

102

bicarbonate, adding to a total of 1.5M Na+. The medium composition was as follows: Na2CO3

103

(33.6 g l-1), NaHCO3 (69.3 g l-1), KHCO3 (1 g l-1), K2HPO4 (1 g l-1), NH4Cl (0.27 g l-1),

104

MgCl2.6H2O (0.1 g l-1), CaCl2.2H2O (0.01 g l-1) and 10 ml l-1 of vitamin solution (12). Two

105

trace element solutions with the following composition were mixed together: TE1 - Na-EDTA

106

(1000 mg l-1), FeCl2.4H20 (370 mg l-1), H3BO3 (60 mg l-1), MnCl2.2H20 (26 mg l-1), 7 ACS Paragon Plus Environment

Environmental Science & Technology

Page 8 of 40

107

CoCl2.6H20 (40 mg l-1), ZnCl2 (10 mg l-1), CuCl2 (3 mg l-1), KAl(SO4)2.12H20 (32 mg l-1),

108

NiCl2.6H20 (31 mg l-1); TE2 – NaOH (40 mg l-1), Na2SiO3.5H2O (10 mg l-1), Na2MoO4.2H2O

109

(10 mg l-1), Na2SeO3.5H2O (10 mg l-1), Na2WO4.2H2O (10 mg l-1). From this mixed solution,

110

10 ml l-1 was added to the medium. As electron acceptor, 3.95 g l-1 (25 mM) of sodium

111

thiosulfate (Na2S2O3) was added.

112

Experimental design

113

The bioreactor was filled with medium and flushed with hydrogen overnight with gas

114

recirculation to lower the redox potential. Afterwards, hydrogen was supplied at 5 ml min-1,

115

the gas recirculation was set to approximately 2.5 l min-1 and the pH was controlled at pH 9 (±

116

0.05). Inoculation of the reactor was defined as time 0. A batch run was performed to start-up

117

the bioreactor and to verify thiosulfate conversion (Start-up, Table 1). Four continuous

118

experiments (Run 1, 2, 3 and 4) with different hydraulic retention times (HRT) were

119

performed as described in Table 1. Another batch experiment (Biofilm run) was performed to

120

investigate the contribution of the reactor wall biofilm to the thiosulfate conversion (Table 1).

121

In the biofilm run, all suspended biomass was removed leaving only the biofilm in the

122

bioreactor and fresh medium was added.

123

Batch bottle experiments

124

To gain further insight into biomass growth and use of H2, batch experiments were performed.

125

The bottles had a total volume of 250 ml. At the beginning of the experiments the bottles

126

contained 100 ml liquid and a headspace of 150 ml. For gas measurements 1 ml headspace

127

was sampled and for liquid composition analysis, 5 ml was collected. To start the experiment,

128

5 ml of biomass obtained at the end of run 4 was inoculated into the serum bottles with sterile

129

medium, pH 9 and 1.5 M Na+, containing 25 mM thiosulfate. Three bottles were incubated

130

with a H2/CO2 (80%/20%) gas phase and three bottles were incubated with a N2/CO2 8 ACS Paragon Plus Environment

Page 9 of 40

Environmental Science & Technology

131

(80%/20%) gas phase. All bottles were incubated at 35oC while shaking at 120 rpm. Chemical

132

controls without biomass were included.

133

Analytical procedures

134

Samples for organic acids were filtered through 0.4 µm pore size filters and analyzed using

135

ultra-high performance liquid chromatography (Dionex ultimate 3000RS, Thermo scientific,

136

Wilmington, MA) as described previously (11). Samples for sulfide, thiosulfate and sulfate

137

analysis were stabilized with zinc acetate (0.2 M) in a 1:1 ratio immediately after sampling.

138

Sulfide was analyzed using a colorimetric test (LCK653, Hach Lange, Düsseldorf, Germany).

139

For sulfate and thiosulfate analysis the samples were filtered through 0.4 µm pore size filters

140

and analyzed by ion chromatography (761 compact IC with a 762 IC interface, Metrohm,

141

Herisau, Switzerland) as described previously (11).

142

Gaseous compounds (H2, CO2, N2 and CH4) were analyzed by gas chromatography using a

143

microGC (CP-4900, Varian, Palo Alto, CA) as described previously (11).

144

The biomass development in the bioreactor was determined by analyzing the total nitrogen

145

(total N) content of biomass. Samples (10 ml) were centrifuged (10 min, 7500g) and washed 3

146

times with carbonate/bicarbonate buffer (pH 9, 1.5M Na+) to remove dissolved nitrogen

147

compounds. The total N content was determined using a colorimetric test (LCK238, Hach

148

Lange, Düsseldorf, Germany).

149

Biomass particle size was quantified using laser measurement in a particle size and shape

150

analyzer (Eyetech, Doner technologies, Or Akiva, Israel) with the Dipa 2000 software (Doner

151

technologies, Or Akiva, Israel). Measurements were done in triplicate for 120s while stirring.

152

Microscopy pictures were taken using a light microscope (DMI6000B, Leica, Biberach,

153

Germany).

9 ACS Paragon Plus Environment

Environmental Science & Technology

154

Page 10 of 40

DNA extraction

155

Samples (50 ml) for DNA analysis were centrifuged (10 min, 7500 g). The biomass was re-

156

suspended in 2 ml of carbonate/bicarbonate buffer (pH 9, 1.5 M Na+) and stored at -80oC.

157

Genomic DNA was extracted from the pellet after centrifugation of stored samples using the

158

PowerBiofilm™ DNA Isolation Kit (MoBio, Carlsbad, CA) following the manufacturer’s

159

instructions. After extraction, the DNA quantity and quality was analyzed using NanoDrop

160

1000 (ThermoScientific, Wilmington, MA). Only samples with more than 5 ng µl-1 DNA and

161

ratio 260/280 nm between 1.7 and 2 were used. When required, the extraction process was

162

repeated to achieve the quantity and quality mentioned. Amplified DNA samples were used

163

for DGGE for a routine check of the microbial community, for a clone library to identify the

164

species in the bioreactor, and for Illumina sequencing for investigating the relative abundance

165

of different OTUs.

166

PCR for DGGE and clone library

167

Bacterial 16S rRNA genes were amplified by PCR using a Taq DNA polymerase kit

168

(Invitrogen, Carlsbad, CA). The primer sets used were U968f/L1401r for denaturing gradient

169

gel electrophoresis (DGGE) and Bact27f/Uni1492r for 16S rRNA cloning and sequencing

170

(13). A 40 bp GC-clamp was added at the 5’ end sequence of the primer U968-f. For archaeal

171

16S rRNA genes the A109(T)f/GC515r primers set were used for DGGE (13, 14).

172

For both bacteria and archaea 16S rRNA gene DGGE, the PCR program was: initial

173

denaturation for 2 min; 35 cycles of 30 s denaturation at 95oC, 40 s at 56oC for annealing and

174

1.5 min elongation step at 72oC; 7 min at 72oC of post-elongation step. PCR settings for

175

bacteria 16S rRNA gene cloning were as described above except that a total number of 25

176

cycles and an annealing temperature of 52oC were used.

177

Clone library 10 ACS Paragon Plus Environment

Page 11 of 40

Environmental Science & Technology

178

The bacterial 16S rRNA amplicons obtained from the final biomass (day 140) were purified

179

using the ssDNA/RNA Clean & Concentrator (Zymo Research, Irvine, CA) and ligated into

180

the pGEM-T Easy Vector System I (Promega, Madison, MI) and cloned in XL-1 blue

181

competent Escherichia coli cells (Stratagene, la Jolla, CA) and grown on LB-agar with 100

182

mg l-1 ampicillin, 0.1 mM isopropyl-1-thio-β-D-galactopyranoside (X-gal). 96 Positive

183

transformants were selected (by blue/white screening) and transferred to solid LB medium

184

with 100 mg l-1 ampicilin and incubated overnight at 37oC. 96 bacterial clones were sent for

185

sequencing of the 16S rRNA gene insert (Baseclear BV, the Netherlands) using the primers

186

sets Bact27f/Univ1492r (Lane, 1991). The DNA sequences were analyzed using Chromas

187

(ver. 2.32, Technelysium). From the 96 clones, 64 good sequences (without chimeras) were

188

obtained and analyzed further. The sequences were in average 1400 base pairs length.

189

Similarity searches for 16S rRNA gene sequences derived from the clones were performed

190

using the EMBL-EBI ENA sequence search program within the ENA database.

191

(http://www.ebi.ac.uk/ena).

192

Bacteria community profiling

193

From the biomass sample on day 140, a fragment of the 16S rRNA gene of bacteria was

194

amplified with primers 341F and 805R (15). The PCR protocol was performed according as

195

previously described (16) and sequenced using the Illumina Miseq platform (17) at the

196

Science for Life Laboratory, Sweden (www.scilifelab.se). The sequence data were processed

197

with the UPARSE pipeline (18) and annotated against the SINA/SILVA database SILVA 119

198

(19). Finally, the data were analyzed using Explicet 2.10.5 (20). Sequences were submitted to

199

the ENA database (http://www.ebi.ac.uk/ena) under the accession number PRJEB11708.

200

Archaea community profiling

11 ACS Paragon Plus Environment

Environmental Science & Technology

Page 12 of 40

201

For Archaea community profiling, an adapted method was used to analyze the biomass

202

sample from day 140 (21). A fragment of the 16S rRNA gene of archaea was amplified with

203

primers 340F and 1000R (22). The PCR protocol was followed using a 50 µl solution with 1

204

µl DNA, 200 nM of the forward and reverse primer, 1 U KOD Hot Start DNA Polymerase

205

(Novagen, Madison, WI), 5 µl KOD-buffer (10x), 3 µl MgSO4 (25 mM), 5 µl dNTP mix (2

206

mM each), and 33 µl sterile water. The PCR program used was: 2 minutes at 95oC and 35

207

cycles of 20s at 95oC, 10s at 5oC and 15s at 70oC. The amplicons were purified with a MSB

208

spin PCRapace kit (Invitek, Dublin, OH) and the concentration measured with a Nanodrop

209

1000 spectrophotometer (Thermo scientific, Wilmington, DE). Equimolar amounts of

210

amplicons from each sample were mixed to a total of 200 ng. This mixed sample was purified

211

with a Purelink PCR Purification kit (Invitrogen, Carlsbad, CA), using a high cutoff binding

212

buffer B3. The samples were sequenced on the 454 Life Sciences GS-FLX platform using

213

Titanium sequencing chemistry (GATC-Biotech, Konstanz, Germany). The sequencing data

214

was analysed using QIIME v1.2 (23) and the chimeric sequences were filtered using Chimera

215

Slayer (24). The OTU clustering was performed as described in QIIME newsletter of

216

December 17th 2010 (http://qiime.wordpress.com/2010/12/17/new-default-parameters-for-

217

uclust-otu-pickers/) and with 97% identity threshold. The diversity metrics were calculated in

218

QIIME 1.2. Hierarchical clustering was done with UPGMA with weighted UniFrac as

219

distance measure in QIIME 1.2 (23). For the taxonomic classification the Ribosomal Database

220

Project (RDP) classifier 2.2 was used (25).

221

Calculations

222

Calculations are described in supplementary material. The following assumptions were made:

223

1. Thermodynamic calculations were made based on Eq. 1 for thiosulfate reduction, Eq. 2 for

224

thiosulfate disproportionation and for sulfate reduction based on Eq. 3.

12 ACS Paragon Plus Environment

Page 13 of 40

225 226

Environmental Science & Technology

   +   + →  +   

(Eq. 3)

2. Evaporation does not cause a major loss of liquid from the bioreactor because it is a closed

227

system. The bioreactor liquid volume is assumed constant during the operation which

228

means that the liquid flow that goes into the bioreactor is equal to the liquid flow getting

229

out of the bioreactor (Qin = Qout).

230

3. Accumulation of sulfur compounds by incorporation in biomass and formation of sulfide

231

precipitates is assumed to play a negligible role due to the high thiosulfate concentration in

232

the influent.

233

4. Hydrogen used for biomass synthesis is assumed to play a negligible role in the bioreactor.

234

5. The N fraction value of 0.2 was used to calculate biomass concentration based on total N,

235 236

following the biomass molecular formula: C1H1.8O0.5N0.2 6. For suspended biomass doubling time, the bioreactor was assumed as continuous stirred

237

tank reactor for biomass growth rate calculations. It was assumed that there is hardly any

238

retention of suspended biomass in the bioreactor. The role of the biomass growing in a

239

biofilm that developed on the glass wall of the bioreactor, in the biomass retention is thus

240

considered negligible for this specific parameter.

241

242

243

Results & Discussion Bioreactor performance

244

The biomass in the bioreactor efficiently converted thiosulfate from the beginning of the

245

operation. A stable maximum volumetric thiosulfate conversion rate of 28.7 (± 0.8) mmolS lr-1

246

d-1 was achieved in run 3 with a hydraulic retention time (HRT) of 1.7 days (Figure 1A and 13 ACS Paragon Plus Environment

Environmental Science & Technology

Page 14 of 40

247

B). This rate is almost two times higher than the sulfate reduction rate in a similar system

248

(Table 2) (11). The higher volumetric rate can be explained by the 2 times higher biomass

249

concentration obtained compared to that in the previous study (Table 2, Figure 1C). Such

250

difference shows that thiosulfate is easier to convert than sulfate which might be related to the

251

energy required for sulfate activation prior to its conversion while thiosulfate does not require

252

activation (26). In addition, there is an intrinsic higher energy gain from thiosulfate reduction

253

to sulfide (∆G’ = -121 kJ per reaction) compared to sulfate reduction (∆G’ = -99 kJ per

254

reaction) at 35oC and pH 9. With a HRT of 1 day (run 4) initially even higher volumetric rates

255

were measured (Figure 1B). Yet, the gradual decrease in biomass concentration and the

256

decrease of volumetric thiosulfate conversion rate showed that the system was not stable at an

257

HRT of 1 day. Apparently, bacteria actively involved in thiosulfate conversion were washed

258

out. This can also be seen in the 16S rRNA gene DGGE bacteria profile where a dominant

259

band disappeared during run 4 (Figure S2). The cause might be that the growth rate of these

260

bacteria was lower than the dilution rate used in run 4, and thus they were washed out of the

261

bioreactor.

262

The loss of bacteria could be prevented by the formation of microbial aggregates which would

263

allow biomass to be retained in the bioreactor. However, microbial aggregation was not

264

observed in the bioreactor even though a 3 phase separator was used to retain microbial

265

aggregates inside the bioreactor. The biomass particle size distribution matches the size of the

266

microorganisms found in the biomass, thick vibrio and thin rods shaped microorganisms,

267

indicative of suspended biomass rather than aggregated biomass (Figure 2). No aggregates

268

were observed by microscopy even though the initial inoculum had small biomass aggregates

269

of around 100 µm in diameter (Figure 2). Absence of biomass aggregation at haloalkaline

270

conditions was also previously observed (11).

271

bioreactor glass wall over time. This biofilm was tested in a batch test (Biofilm run) to

However, a biofilm developed on the

14 ACS Paragon Plus Environment

Page 15 of 40

Environmental Science & Technology

272

determine its contribution to the bioreactor performance. In this run, not only thiosulfate

273

conversion and formate production, but also acetate and methane production was observed

274

(Figure 3). Acetate and methane were not detected during continuous operation. Acetate, if

275

intermediately produced during continuous operation, might also have been used as carbon

276

source by heterotrophic sulfidogenic microorganisms. In the biomass, bacteria were detected

277

that are phylogenetically closely related to sulfate/thiosulfate converting bacteria that use

278

acetate as carbon source (Table 3). Some Desulfonatronovibrio sp. are unable to grow

279

autotrophically with H2 and formate, but are able to grow with acetate carbon source and

280

hydrogen or formate as electron donor (27, 28).

281

Thiosulfate conversion

282

Thiosulfate conversion occurred either through disproportionation and subsequent reduction

283

of sulfate or through direct reduction (Figure 1A). In previous studies with marine sediments

284

thiosulfate conversion was shown to occur partly by disproportionation and partly by

285

reduction (1). Disproportionation occurred as sulfate was produced in the bioreactor. The

286

production of sulfate from thiosulfate using O2 or nitrate as electron acceptor can be discarded

287

since these were not present in our system (8). Phototrophic microorganisms can also produce

288

sulfate from thiosulfate but such microorganisms were not found in microbial community

289

analysis (Table 3), and therefore cannot explain sulfate production (8). Thus, thiosulfate

290

disproportionation is the most probable metabolism for sulfate production. Thiosulfate

291

disproportionation can occur biologically and chemically (3, 29). However, chemical

292

disproportionation did not occur at the conditions used as can be seen in the batch bottles

293

experiments (Figure 4C). Clones with similarity to known thiosulfate-disproportionating

294

bacteria (Desulfonatronovibrio and Desulfonatronospira) were identified (Table 3).

295

Desulfonatronovibrio and Desulfonatronospira are also capable of sulfate reduction, which

296

was also observed. For the two isolated and characterized species it was found that 15 ACS Paragon Plus Environment

Environmental Science & Technology

Page 16 of 40

297

disproportionation can occur without electron donor (27). In the current study, thiosulfate

298

disproportionation also took place when electron donors were available (H2 supplied and

299

formate produced (Figure S3). Assuming a complete mixing of the bioreactor, it is unlikely

300

that there were regions where biomass did not have access to at least one of these electron

301

donors.

302

To better understand the effect of H2 and formate on the overall thiosulfate conversion, batch

303

experiments were performed. When no electron donor was available, only thiosulfate

304

disproportionation occurred (Figure 4B). When H2 and/or formate were available, thiosulfate

305

was mainly converted to sulfide and some sulfate, confirming that disproportionation

306

occurred in the presence of an electron donor (Figure 4A). When all electron donors were

307

depleted, after day 9, sulfate accumulated due to disproportionation (Figure 4A, 4D and 4E).

308

These results reinforce the idea that thiosulfate disproportionation occurs even in the presence

309

of an electron donor. More detailed studies are required to understand the mechanism of

310

thiosulfate disproportionation.

311

Microbial community

312

The microbial community in the bioreactor changed when the HRT was lowered during the

313

bioreactor operation and was different between suspended biomass and biomass attached to

314

the bioreactor walls (Figure S2).

315

For bacteria, there is a clear decrease in diversity with the decrease of HRT. The bacteria with

316

higher growth rates are kept in the bioreactor while the ones that have growth rates lower than

317

dilution rate get washed out, unless they grow attached to the bioreactor walls. This also

318

suggests that adding support material to the bioreactor might improve the thiosulfate

319

conversion efficiencies by keeping more biomass and microbial diversity inside the

320

bioreactor. 16 ACS Paragon Plus Environment

Page 17 of 40

Environmental Science & Technology

321

A decrease in archaeal diversity with lower HRT was not observed (Figure S2). The archaeal

322

community of suspended and attached biomass was similar, which indicates that the archaea

323

in the bioreactor were not affected by the dilution rate applied.

324

The microbial community obtained at the end of run 4 (day 140) was studied further. A 16S

325

rRNA gene clone library for bacteria was made and used to identify phylogenetic affiliation

326

of the dominant bacteria while a 16S rRNA gene MiSeq approach for both bacteria and

327

archaea was used to study the relative abundance of different microorganisms (Table 3;

328

Figure 5). Bacteria belonging to the family Clostridiaceae were dominant in the bioreactor

329

(Figure 5). In the clone library, 16S rRNA genes that have 97% similarity to Tindallia

330

texcoconensis were identified (Table 3). These clones were also dominant in a hydrogen fed

331

sulfate-reducing bioreactor (11). Tindallia texcoconensis is a fermentative bacterium that uses

332

peptone and a few amino acids for growth (30). In our study, however, no organic compounds

333

were fed to the bioreactor. Due to the low concentration of biomass in the bioreactor, it is very

334

unlikely that such microorganisms can become dominant by feeding on dead biomass alone.

335

Another isolate that is closely related to Tindallia magadiensis was able to produce formate

336

from H2 and CO2 and use it to reduce thiosulfate (4). In our bioreactor, high concentrations of

337

formate (up to 100 mM) were found that could have been produced by the microorganisms

338

closely related to Tindallia sp. (Figure S3). At the actual conditions in the bioreactor: pH 9, 1

339

atm of H2 and 0.825 M of HCO3-, formate production from hydrogen (Eq. 4) can yield energy.

340

The Gibbs free energy is -24.3 kJ when formate concentration is low, 0.01 mM, and is close

341

to the thermodynamic equilibrium (6.88 kJ) when formate reaches 90 mM.

342

 +  →   +  

343

Thiosulfate-reducing and thiosulfate-disproportionating bacteria, belonging to the family

344

Desulfohalobiaceae, only represent a small fraction of the total population, approximately

(Eq. 4)

17 ACS Paragon Plus Environment

Environmental Science & Technology

Page 18 of 40

345

10% (Figure 5). From this family Desulfonatronovibrio sp. and Desulfonatronospira

346

thiodismutans related clones were identified (Table 3). These two genera can reduce sulfate.

347

Sequences related to other known thiosulfate-reducing bacteria were identified, such as

348

Desulfurivibrio sp. and Dethiobacter sp. (Table 3). The species previously studied that belong

349

to these genera cannot use sulfate as electron acceptor. Likely, their role in the bioreactor was

350

restricted to thiosulfate disproportionation or thiosulfate reduction.

351

Surprisingly,

352

Halothiobacillaceae represented approximately 10-14% and 3-6% of the total microbial

353

community, respectively (Figure 5). In the clone library 28 out of 64 clones were closely

354

related to Thiomicrospira sp. from the family Piscirickettsiaceae, and Thioalkalibacter sp.

355

from the family Halothiobacillacea. These bacteria are mainly known as sulfur-oxidizing

356

bacteria that thrive in oxic and micro-oxic conditions (31). As O2 and nitrate were not added

357

to the bioreactor the physiological role of these microorganisms remains unknown. Sulfide

358

oxidizers were previously also identified in sulfate-reducing bioreactors operated at

359

haloalkaline conditions, where their role was not established (32). It seems plausible to

360

assume that sulfide-oxidizing microorganisms might be capable of disproportionation of

361

sulfur compounds such as thiosulfate (33). However, this was never showed for

362

haloalkaliphilc sulfide oxidizing microorganisms.

363

As methane production occurred in the biofilm run, the archaeal community was also

364

investigated in the biomass at the end of the reactor run. Archaea belonging to the genus

365

Methanocalculus were most dominant, approximately 82% of the total archaea.

366

Methanocalculus sp. are known as haloalkaliphilic methanogens that can use H2/CO2 or

367

formate as growth substrates (34).

368

microorganisms

related

to

the

family

Piscirickettsiaceae

and

Application outlook

18 ACS Paragon Plus Environment

Page 19 of 40

Environmental Science & Technology

369

The results clearly show that it is possible to efficiently convert thiosulfate to sulfide using a

370

bioreactor fed with hydrogen as electron donor and without any addition of organic

371

compounds. The conversion of thiosulfate to sulfide can thus be used in biodesulfurization

372

processes to treat the bleed stream and reduced the water and caustic requirements. Two

373

different approaches for a practical application of thiosulfate conversion can be considered

374

(Table 4). One approach is using H2 or formate as electron donor to fully reduced thiosulfate

375

to sulfide and eliminate the need for bleed stream disposal. Another new approach is without

376

addition of electron donor by making use of thiosulfate disproportionation to sulfate and

377

sulfide. The option without electron donor could be applied to selectively remove thiosulfate

378

from biodesulfurization bleed streams without the additional cost of an electron donor (Figure

379

6). By disproportionating thiosulfate to sulfate and sulfide and recycling the bleed stream

380

back, sulfate would become the main component of the bleed stream. Thiosulfate adds to the

381

total COD of the bleed stream while sulfate does not. Due to COD disposal restrictions,

382

removing thiosulfate without use of electron donor might be an attractive option to decrease

383

the bleed stream COD content.

384

Thiosulfate disproportionation results in a low amount of energy per reaction. It yields -22 kJ

385

at standard conditions (pH 7 and 25oC) and -62 kJ when taking into account the in situ

386

concentrations, pH 9 and 25 mM thiosulfate. This limits biomass growth in a bioreactor

387

without electron donor. But, at the conditions used in the bioreactor, pH 9 and about 1.5M

388

sodium carbonate/bicarbonate, thiosulfate disproportionation yields more energy, -89 kJ per

389

reaction.

390

thermodynamically more favourable at high pH. Biomass growth was determined in small

391

batch tests. It became clear that microorganisms grow by thiosulfate disproportionation, even

392

though the growth yield is less than with electron donor (Figure S4). Thiosulfate

Because

the

disproportionation

produces

protons,

it

actually

becomes

19 ACS Paragon Plus Environment

Environmental Science & Technology

Page 20 of 40

393

disproportionation should be further explored in bioreactor systems to get insight into its

394

feasibility for future large scale application.

395

396

Acknowledgments

397

This work was performed in the TTIW-cooperation framework of Wetsus, European Centre

398

of Excellence for Sustainable Water Technology (www.wetsus.nl). Wetsus is funded by the

399

Dutch Ministry of Economic Affairs, the European Union Regional Development Fund, the

400

Province of Fryslân, the City of Leeuwarden and the EZ/Kompas program of the

401

“Samenwerkingsverband Noord-Nederland”. The authors thank the participants of the

402

research theme “Sulfur”, namely Paqell, for fruitful discussions and financial support.

403

Research of AJMS is financed by ERC grant project 323009 and by Gravitation grant project

404

024.002.002 from the Netherlands Ministry of Education, Culture and Science. Stephan

405

Christel and Mark Dopson (Centre for Ecology and Evolution in Microbial Model Systems,

406

Linnaeus University, Sweden) are gratefully acknowledged for the NGS of the bacteria. There

407

is no conflict of interest to declare related to this work.

408

409

Supporting Information Available

410

The following supporting information is available for this manuscript: figure S1, figure S2,

411

figure S3, figure S4 and calculations. This information is available free of charge via the

412

Internet at http://pubs.acs.org.

413

414

References 20 ACS Paragon Plus Environment

Page 21 of 40

Environmental Science & Technology

415 (1)

Jørgensen, B. B. 1990. A thiosulfate shunt in the sulfur cycle of marine sediments.

416

Science. 1990, 249, 152-154; DOI 10.1126/science.249.4965.152.

417 (2)

Widdel, F.; Microbiology and ecology of sulfate- and sulfur-reducing bacteria. In: The

418

Biology of Anaerobic Microorganisms; Zehnder, A. J. B., Ed.; Wiley: New York 1988;

419

pp. 469-586.

420 (3)

Bak, F.; Cypionka, H. A novel type of energy metabolism involving fermentation of

421

inorganic sulphur compounds. Nature. 1987, 326, 891-892; DOI 10.1038/326891a0.

422 (4)

Sorokin, D. Y.; Detkova, E. N.; Muyzer, G. Sulfur-dependent respiration under extremely

423

haloalkaline conditions in soda lake “acetogens” and the description of Natroniella

424

sulfidigena sp. nov. FEMS Microbiol. Lett. 2011, 319, 88-95; DOI 10.1111/j.1574-

425

6968.2011.02272.x.

426 (5)

Finster, K. Microbiological disproportionation of inorganic sulfur compounds. J. Sulfur

427

Chem. 2008, 29, 37-41; DOI 10.1080/17415990802105770

428 (6)

Van den Bosch, P. L. F.; Sorokin, D. Y.; Buisman, C. J. N.; Janssen, A. J. H. The effect

429

of pH on thiosulfate formation in a biotechnological process for the removal of hydrogen

430

sulfide from gas streams. Environ. Sci. Technol. 2008, 42, 2637-2642; DOI

431

10.1021/es7024438.

432 (7)

Sorokin, D. Y.; Rusanov, I. I.; Pimenov, N. V.; Tourova, T. P.; Abbas, B.; Muyzer, G.

433

Sulfidogenesis under extremely haloalkaline conditions in soda lakes of Kulunda Steppe

434

(Altai, Russia). FEMS Microbiol. Ecol. 2010, 73, 278-290; DOI 10.1111/j.1574-

435

6941.2010.00901.x.

436 (8)

Sorokin, D. Y.; Kuenen, J. G.; Muyzer, G. The microbial sulfur cycle at extremely

437

haloalkaline conditions of soda lakes. Front. Microbiol. 2011, 2, 1-16; DOI

438

10.3389/fmicb.2011.00044.

21 ACS Paragon Plus Environment

Environmental Science & Technology

Page 22 of 40

439 (9)

Janssen, A. J. H.; Lens, P. N. L.; Stams, A. J. M.; Plugge, C. M.; Sorokin, D. Y.; Muyzer,

440

G.; Dijkman, H.; Van Zessen, E.; Luimes, P.; Buisman, C. J. N. Application of bacteria

441

involved in the biological sulfur cycle for paper mill effluent purification. Sci. Total

442

Environ. 2009, 407, 1333-1343; DOI 10.1016/j.scitotenv.2008.09.054.

443 (10) Buisman, C. J. N.; Post, R.; IJspeert, P.; Geraats, G.; Lettinga, G. Biotechnological 444

process for sulphide removal with sulphur reclamation. Acta Biotechnol. 1989, 9, 255-

445

267; DOI 10.1002/abio.370090313.

446 (11) Sousa, J. A. B.; Plugge, C. M.; Stams, A. J. M.; Bijmans, M. F. M. Sulfate reduction in a 447

hydrogen fed bioreactor operated at haloalkaline conditions. Water Res. 2015, 68, 67-76;

448

DOI 10.1016/j.watres.2014.09.035

449 (12) Wolin, E. A.; Wolin, M. J.; Wolfe, R. S. Formation of methane by bacterial extract. J. 450

Biol. Chem. 1963, 238, 2882-2886.

451 (13) Lane, D.L.; 16S/23S rRNA sequencing. In: Nucleic Acid Techniques in Bacterial 452

Systematics; Stackebrandt E., Goodfellow M., Eds.; Nucleic Acid Techniques in Bacterial

453

Systematics; John Wiley & Sons Ltd: Chichester, United Kingdom 1991; pp. 115-175.

454 (14) Großkopf, R.; Janssen, P.H.; Liesack, W. Diversity and structure of the methanogenic 455

community in anoxic rice paddy soil microcosms as examined by cultivation and direct

456

16S rRNA gene sequence retrieval. Appl. Environ. Microbiol. 1998, 64, 960-969; DOI

457

0099-2240/98/$04.0010.

458 (15) Herlemann, D.P.; Labrenz, M.; Jurgens, K.; Bertilsson, S.; Waniek, J.J.; Andersson, A.F. 459

Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic

460

Sea. ISME J. 2011, 5, 1571-1579; DOI 10.1038/ismej.2011.41.

461 (16) Hugerth, L.W.; Wefer, H.A.; Lundin, S.; Jakobsson, H.E.; Lindberg, M.; Rodin, S.; 462

Engstrand, L.; Andersson, A.F. DegePrime, a program for degenerate primer design for

22 ACS Paragon Plus Environment

Page 23 of 40

Environmental Science & Technology

463

broad-taxonomic-range PCR in microbial ecology studies. Appl. Environ. Microbiol.

464

2014, 80, 5116-5123; DOI 10.1128/AEM.01403-14.

465 (17) Lindh, M.V.; Figueroa, D.; Sjöstedt, J.; Baltar, F.; Lundin, D.; Andersson, A.; Legrand, 466

C.; Pinhassi, J. Transplant experiments uncover Baltic Sea basin-specific responses in

467

bacterioplankton community composition and metabolic activities. Front. Microbiol.

468

2015, 6, 223; DOI 10.3389/fmicb.2015.00223.

469 (18) Edgar, R.C. UPARSE: highly accurate OTU sequences from microbial amplicon reads. 470

Nature Methods. 2013, 10, 996-998; DOI 10.1038/nmeth.2604.

471 (19) Quast, C.; Pruesse, E.; Yilmaz, P.; Gerken, J.; Schweer, T.; Yarza, P.; Peplies, J.; 472

Glöckner, F.O. The SILVA ribosomal RNA gene database project: improved data

473

processing and web-based tools. Nucl. Acids Res. 2013, 41, 590-596; DOI

474

10.1093/nar/gks1219.

475 (20) Robertson, C.E.; Harris, J.K.; Wagner, B.D.; Granger, D.; Browne, K.; Tatem, B.; Feazel, 476

L.M.; Park, K.; Pace, N.R.; Frank, D.N. Explicet: graphical user interface software for

477

metadata-driven management, analysis and visualization of microbiome data.

478

Bioinformatics. 2013, 29, 3100-3101; DOI 10.1093/bioinformatics/btt526.

479 (21) Jaeggi, T.; Kortman, G.A.; Moretti, D.; Chassard, C.; Holding, P.; Dostal, A.; Boekhorst, 480

J.; Timmerman, H.M.; Swinkels, D.W.; Tjalsma, H.; Njenga, J.; Mwangi, A.; Kvalsvig,

481

J.; Lacroix, C.; Zimmermann, M.B. Iron fortification adversely affects the gut

482

microbiome, increases pathogen abundance and induces intestinal inflammation in

483

Kenyan infants. Gut. 2014, 64, 731-742; DOI 10.1136/gutjnl-2014-307720.

484 (22) Gantner, S.A.; Andersson, F.; Alonso-Saez, L.; Bertilsson, S. Novel primers for 16S 485

rRNA-based archaeal community analyses in environmental samples. J Microbiol.

486

Methods. 2011, 84, 12-18; DOI 10.1016/j.mimet.2010.10.001.

23 ACS Paragon Plus Environment

Environmental Science & Technology

Page 24 of 40

487 (23) Caporaso, J.G.; Kuczynski, J.; Stombaugh, J.; Bittinger, K.; Bushman, F.D.; Costello, 488

E.K.; Fierer, N.; Pena, A.G.; Goodrich, J.K.; Gordon, J.I.; Huttley, G.A.; Kelley, S.T.;

489

Knights, D.; Koenig, J.E.; Ley, R.E.; Lozupone, C.A.; McDonald, D.; Muegge, B.D.;

490

Pirrung, M.; Reeder, J.; Sevinsky, J.R.; Turnbaugh, P.J.; Walters, W.A.; Widmann, J.;

491

Yatsunenko, T.; Zaneveld, J.; Knight, R. QIIME allows analysis of high-throughput

492

community sequencing data. Nature Methods. 2010, 7, 335-336; DOI

493

10.1038/nmeth.f.303.

494 (24) Haas, B.J.; Gevers, D.; Earl, A.M.; Feldgarden, M.; Ward, D.V.; Giannoukos, G.; Ciulla, 495

D.; Tabbaa, D.; Highlander, S.K.; Sodergren, E.; Methe, B.; DeSantis, T.Z.; Human

496

Microbiome, C.; Petrosino, J.F.; Knight, R.; Birren, B.W. Chimeric 16S rRNA sequence

497

formation and detection in Sanger and 454-pyrosequenced PCR amplicons. Genome Res.

498

2011, 21, 494-504; DOI 10.1101/gr.112730.110.

499 (25) Cole, J.R.; Wang, Q.; Cardenas, E.; Fish, J.; Chai, B.; Farris, R.J.; Kulam-Syed500

Mohideen, A.S.; McGarrell, D.M.; Marsh, T.; Garrit, G.M.; Tiedje, J.M. The Ribosomal

501

Database Project: improved alignments and new tools for rRNA analysis. Nucl. Acids

502

Res. 2009, 37, 141-145; DOI 10.1093/nar/gkn879.

503 (26) Sydow, U.; Wohland, P.; Wolke, I.; Cypionka, H. Bioenergetics of the alkaliphilic 504

sulfate-reducing bacterium Desulfonatronovibrio hydrogenovorans. Microbiol. 2002,

505

148, 853-860; DOI 10.1099/00221287-148-3-853

506 (27) Sorokin, D.Y.; Tourova, T.P.; Kolganova, T.V.; Detkova, E.N.; Galinski, E.A.; Muyzer, 507

G. Culturable diversity of lithotrophic haloalkaliphilic sulfate-reducing bacteria in soda

508

lakes and the description of Desulfonatronum thioautotrophicum sp. nov.,

509

Desulfonatronum thiosulfatophilum sp. nov., and Desulfonatronovibrio thiodismutans sp.

510

nov., and Desulfonatronovibrio magnus sp. nov.. Extremophiles. 2011, 15, 391-401; DOI

511

10.1007/s00792-011-0370-7.

24 ACS Paragon Plus Environment

Page 25 of 40

Environmental Science & Technology

512 (28) Sorokin, D.Y.; Tourova, T.P.; Henstra, A.M.; Stams, A.J.M.; Galinski, E.A.; Muyzer, G. 513

Sulfidogenesis under extremely haloalkaline conditions by Desulfonatronospira

514

thiodismutans gen. nov., sp. nov., and Desulfonatronospira delicata sp. nov. - a novel

515

lineage of Deltaproteobacteria from hypersaline soda lakes. Microbiol. 2008, 154, 1444-

516

1453; DOI 10.1099/mic.0.2007/015628-0.

517 (29) Kamyshny, A.; Druschel, G.; Mansaray, Z.F.; Farquhar, J. Multiple sulfur isotopes 518

fractionations associated with abiotic sulfur transformations in Yellowstone National

519

Park geothermal springs. Geochem. Transactions. 2014, 15, 1-22; DOI 10.1186/1467-

520

4866-15-7.

521 (30) Alazard, D.; Badillo, C.; Fardeau, M.L.; Cayol, J.L.; Thomas, P.; Roldan, T.; Tholozan, 522

J.L.; Ollivier, B. Tindallia texcoconensis sp. nov., a new haloalkaliphilic bacterium

523

isolated from lake Texcoco, Mexico. Extremophiles. 2006, 11, 33-39; DOI

524

10.1007/s00792-006-0006-5.

525 (31) Sorokin, D.Y.; Tourova, T.P.; Lysenko, A.M.; Muyzer, G. Diversity of culturable 526

halophilic sulfur-oxidizing bacteria in hypersaline habitats. Microbiol. 2006, 152, 3013-

527

3023; DOI 10.1099/mic.0.29106-0.

528 (32) Zhou, J.; Zhou, X.; Li, Y.; Xing, J. Bacterial communities in haloalkaliphilic sulfate529

reducing bioreactors under different electron donors revealed by 16S rRNA MiSeq

530

sequencing. J. Hazard. Mater. 2015, 295, 176-184; DOI 10.1016/j.jhazmat.2015.04.010.

531

(33) Warren, L.A.; Norlund, L.I.; Bernier, L. Microbial thiosulphate reaction arrays: the

532

interactive roles of Fe(III), O2 and microbial strain on disproportionation and oxidation

533

pathways. Geobiol. 2008, 6, 461-470; DOI 10.1111/j.1472-4669.2008.00173.x

534 (34) Sorokin, D.Y.; Abbas, B.; Geleijnse, M.; Pimenov, N.V.; Sukhacheva, M.V.; van 535

Loosdrecht, M.C.M. Methanogenesis at extremely haloalkaline conditions in the soda

25 ACS Paragon Plus Environment

Environmental Science & Technology

536

lakes of Kulunda Steppe (Altai, Russia). FEMS Microbiol. Ecol. 2015, 91, 16; DOI

537

10.1093/femsec/fiv016.

Page 26 of 40

538 (35) Zhou, J.; Xing, J. Effect of electron donors on the performance of haloalkaliphilic sulfate539

reducing bioreactors for flue gas treatment and microbial degradation patterns related to

540

sulfate reduction of different electron donors. Biochem. Eng. J. 2015, 96, 14-22; DOI

541

10.1016/j.bej.2014.12.015.

542

543

26 ACS Paragon Plus Environment

Page 27 of 40

544

Environmental Science & Technology

Captions

545 546

Figure 1. Performance during start-up and continuous run of the H2 fed thiosulfate converting

547

gas-lift bioreactor operated at haloalkaline conditions. A – Thiosulfate, sulfate and sulfide

548

concentrations in the bioreactor effluent. B – Thiosulfate loading rate and thiosulfate

549

conversion volumetric rate. C – Biomass concentration (Cx) in the bioreactor.

550 551

Figure 2. Biomass micrograph and particle size distribution. A – Micrograph with the

552

different microbial morphotypes found. White bar represents 10 µm. B – Particle size

553

distribution by biomass volume using laser measurement (3 measurements of 120s with

554

mixing) of biomass from day 0 and 140 of continuous runs (begin of Start-up and end of Run

555

4).

556 557

Figure 3. Thiosulfate, sulfate, sulfide, formate and acetate concentration in the liquid phase

558

and methane fraction in the gas phase during the Biofilm run.

559 560

Figure 4. Thiosulfate conversion and H2 use in batch tests. A – Average thiosulfate (●),

561

sulfate (○) and sulfide (▼) concentration with H2. B – Average thiosulfate (●), sulfate (○),

562

sulfide (▼) concentration with N2. C - Average thiosulfate (●), sulfate (○), sulfide (▼)

563

concentration in the chemical controls with N2. D – Average formate (■) concentration with

564

H2. E – Average H2 (◊) and CH4 (♦) concentration in the gas phase with H2.

565 566

Figure 5. Bacteria and Archaea 16S rRNA relative abundance at the end of Run 4 on family

567

level. OTUs with less than 0.5% relative abundance were grouped in “others”. 27 ACS Paragon Plus Environment

Environmental Science & Technology

Page 28 of 40

568

Figure 6. Biological process for gas desulfurization using an anaerobic thiosulfate (S2O32-)

569

disproportionating bioreactor to reduce the bleed stream COD content. The sulfide (H2S)

570

present in the gas is solubilized in an alkaline solution as HS- using a scrubber (1). This HS-

571

rich solution goes to an aerobic bioreactor (2) where is biologically oxidized, under controlled

572

microaerophilic conditions, to mostly elemental sulfur (S0) and in small part to sulfate (SO42-)

573

and also chemically oxidized, via polysulfides, to thiosulfate (S2O32-). The S0 is separated in a

574

settler (3) and most of the liquid is recycled to the scrubber (1) to solubilize more H2S. Other

575

part of the liquid goes through an anaerobic bioreactor (4) where S2O32-, without addition of

576

e-donor, is disproportionated to SO42- and HS- which are recycled back to the aerobic

577

bioreactor (2). With time this prevents accumulation of S2O32- in the whole system. When a

578

bleed stream is discarded to add new caustic to the system, this bleed stream will have low

579

chemical oxygen demand (COD) because only S2O32- and HS- contribute to the COD and not

580

SO42-.

581

28 ACS Paragon Plus Environment

Page 29 of 40

Environmental Science & Technology

Table 1. Operational differences between the different bioreactor runs performed Period

Mode

(d)

S2O32- loading rate

(d)

(mmol d-1)

Start-up

0-12

Batch

N.A.a

N.A.a

Run 1

12-42

Continuous

5

10

Run 2

42-91

Continuous

2.5

20

Run 3

91-117

Continuous

1.7

29.4

Run 4

117-140

Continuous

1

50

13b

Batch

N.A.a

N.A.a

Biofilm run

582 583 584

HRT

a b

Not applicable. Duration of the batch test.

585

586

587

588

589

590

591

592

29 ACS Paragon Plus Environment

Environmental Science & Technology

Page 30 of 40

Table 2. Conditions and performance of sulfate and thiosulfate reducing bioreactors operated at haloalkaline conditions. This study

(11)a

(35)a

(35)a

Gas lift with 3 phase separator

Gas lift with 3 phase separator

Anaerobic filter

Anaerobic filter

Thiosulfate

Sulfate

Sulfate

Sulfate

H2

H2

Formate

Ethanol

CO2

CO2

Formate

Ethanol

9

9

9.5

9.5

Na+ conc. (M)

1.5

1.5

1

1

Temperature (oC)

35

35

37

37

HRT (d)

1.7

3.3

1

1

28,7 (± 0.8)

18

85

89.5

Formate

Formate

Acetate

Acetate/Formate/ Lactate

14.0 (± 2.2)

7.2 (± 3)

N.D.b

N.D.b

No aggregation

No aggregation

N.D.b

N.D.b

Reactor type e- acceptor e- donor Carbon source pH

Sulfate/Thiosulfate conversion rates (mmolS lr-1 d-1) Side products Biomass conc. (mg l-1) Biomass aggregation

593

a

Data from reference.

594

b

No data available.

595 596 597 598 599 600 601 602 30 ACS Paragon Plus Environment

Page 31 of 40

Environmental Science & Technology

Table 3. Phylogenetic affiliations and frequency of cloned bacterial 16S rRNA gene amplicons retrieved from the bioreactor at day 140. Similarity Closest cultured relative

Thioalkalibacter halophilus

(%)

Number of clones

Metabolism described in

Accession

literature

number

(64 clones)

99

14

Sulfur oxidation

NR 044406

99

14

Sulfur oxidation

AB634592

97

11

Fermentation

NR 043664

99

4

Sulfur oxidation

NR 024854

strain ALCO1 Thiomicrospira sp. V2501 Tindallia texcoconensis strain IMP-300 Thiomicrospira thyasirae strain DSM 5322 Desulfonatronovibrio sp.

Sulfate reduction & 99

4

AHT22 Candidate division SR1

thiosulfate

GU196831

disproportionation 85

4

Halomonas sp. IB-559

99

2

Desulfurivibrio sp. AMeS2

96

2

Sulfur disproportionation

KF148062

Bacteroidetes VNs52

77

2

N.D.*

FJ168485

bacterium RAAC1_SR1_1

Desulfonatronospira

Heterotrophic sulfide oxidation

CP006913

AJ309560

Sulfate reduction & 98

1

thiodismutans ASO3-1 Desulfonatronovibrio sp.

Fermentation

thiosulfate

NR 044459

disproportionation 97

1

Thiosulfate reduction

FJ469580

99

1

Sulfur oxidation

AF013976

89

1

86

1

94

1

Iron reduction

DQ631799

92

1

Fermentation

NR 044361

AHT 10 Thiomicrospira sp. JB-A1F Dethiobacter alkaliphilus AHT 1 Bacillus sp. IST-38 Low G+C Gram-positive

Thiosulfate reduction & Sulfur disproportionation Fermentation & iron reduction

NR 044205

FM877978

bacterium IRB 1 Gracilimonas tropica strain CL-CB462

603

16S rRNA gene sequences were deposited in GenBank with the accession numbers KT990232 – KT990304

31 ACS Paragon Plus Environment

Environmental Science & Technology

604

Page 32 of 40

*No data available.

605

32 ACS Paragon Plus Environment

Page 33 of 40

Environmental Science & Technology

606

Table 4. Relative comparison of three biodesulfurization processes. Process

Caustic

Volume of

Bleed stream

Additional chemical

consumption

Waste bleed

neutralization

consumption

stream disposal Standard biodesulfurization

High

High

COD and Sulfate

None

c,d

removal required With thiosulfate completely reduced to sulfide With thiosulfate disproportionationb

607 608 609 610 611 612 613

Low

Low

Not required

Electron donor (e.g. H2)

Medium

Medium

Sulfate removal

None or very lowe

a

requiredd

a

Described in this work. Proposed in this work. c COD (mainly from thiosulfate) removal is usually performed by aerobic biologic treatment. d Sulfate removal is usually performed by precipitation with lime. e An organic carbon source may be beneficial to support biomass growth. b

614

615

616

617

618

619

33 ACS Paragon Plus Environment

Environmental Science & Technology

Figure 1. Performance during start-up and continuous run of the H2 fed thiosulfate converting gas-lift bioreactor operated at haloalkaline conditions. A – Thiosulfate, sulfate and sulfide concentrations in the bioreactor effluent. B – Thiosulfate loading rate and thiosulfate conversion volumetric rate. C – Biomass concentration (Cx) in the bioreactor. 154x274mm (600 x 600 DPI)

ACS Paragon Plus Environment

Page 34 of 40

Page 35 of 40

Environmental Science & Technology

Figure 2. Biomass micrograph and particle size distribution. A – Micrograph with the different microbial morphotypes found. White bar represents 10 µm. B – Particle size distribution by biomass volume using laser measurement (3 measurements of 120s with mixing) of biomass from day 0 and 140 of continuous runs (begin of Start-up and end of Run 4). 118x161mm (600 x 600 DPI)

ACS Paragon Plus Environment

Environmental Science & Technology

Figure 3. Thiosulfate, sulfate, sulfide, formate and acetate concentration in the liquid phase and methane fraction in the gas phase during the Biofilm run. 122x171mm (600 x 600 DPI)

ACS Paragon Plus Environment

Page 36 of 40

Page 37 of 40

Environmental Science & Technology

Figure 4. Thiosulfate conversion and H2 use in batch tests. A – Average thiosulfate (●), sulfate (○) and sulfide (▼) concentration with H2. B – Average thiosulfate (●), sulfate (○), sulfide (▼) concentration with N2. C - Average thiosulfate (●), sulfate (○), sulfide (▼) concentration in the chemical controls with N2. D – Average formate (■) concentration with H2. E – Average H2 (◊) and CH4 (♦) concentration in the gas phase with H2. 120x137mm (600 x 600 DPI)

ACS Paragon Plus Environment

Environmental Science & Technology

Figure 5. Bacteria and Archaea 16S rRNA relative abundance at the end of Run 4 on family level. OTUs with less than 0.5% relative abundance were grouped in “others”. 88x44mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 38 of 40

Page 39 of 40

Environmental Science & Technology

Figure 6. Biological process for gas desulfurization using an anaerobic thiosulfate (S2O32-) disproportionating bioreactor to reduce the bleed stream COD content. The sulfide (H2S) present in the gas is solubilized in an alkaline solution as HS- using a scrubber (1). This HS- rich solution goes to an aerobic bioreactor (2) where is biologically oxidized, under controlled microaerophilic conditions, to mostly elemental sulfur (S0) and in small part to sulfate (SO42-) and also chemically oxidized, via polysulfides, to thiosulfate (S2O32-). The S0 is separated in a settler (3) and most of the liquid is recycled to the scrubber (1) to solubilize more H2S. Other part of the liquid goes through an anaerobic bioreactor (4) where S2O32-, without addition of e-donor, is disproportionated to SO42- and HS- which are recycled back to the aerobic bioreactor (2). With time this prevents accumulation of S2O32- in the whole system. When a bleed stream is discarded to add new caustic to the system, this bleed stream will have low chemical oxygen demand (COD) because only S2O32- and HS- contribute to the COD and not SO42-. 114x73mm (300 x 300 DPI)

ACS Paragon Plus Environment

Environmental Science & Technology

Graphical abstract 147x109mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 40 of 40