Enhanced hydrogen production by photofermentative microbial aggregation induced by L-cysteine: the effect of substrate concentration, C/N ratio and agitation rate

Abstract

Continuous biomass washout from photobioreactors resulting from the poor flocculation of photofermentative bacteria (PFB) diverts electron flux away from hydrogen production to continuing biosynthesis. This work showed that three key parameters, the substrate concentration, C/N ratio and agitation rate significantly influenced the flocculation behavior of Rhodopseudomonas faecalis RLD-53. Excessive substrate (75 mmol l⁻¹) led an apparent drop in the flocculability, while a proper content of extracellular polymeric substances (EPS) at a C/N ratio of 10 : 2 was beneficial to aggregation of R. faecalis RLD-53. Control of the appropriate agitation rate not only enhanced EPS production and aggregation of PFB, but also avoided disruption of PFB aggregation caused by intensive hydrodynamic shear forces generated from excessive agitation. The optimal flocculability (47.67%) of R. faecalis RLD-53 was obtained at a substrate concentration of 50 mmol l⁻¹, C/N ratio of 10 : 2 and agitation rate of 80 rpm. In the entire continuous operation, most PFB aggregates accumulated in the photobioreactor with only 0.55 g of biomass washout, much less than 2.86 g from free cell culture. Consequently, the maximum hydrogen yield of 2.57 mol H₂ per mol acetate and production rate of 706 ml l⁻¹ d⁻¹ were obtained by PFB aggregates, as biomass synthesis was greatly reduced. Therefore, aggregation of PFB is a promising strategy to mediate the reductant away from biomass synthesis into hydrogen production in continuous operation.

---

1. Introduction

Environmental pollution and climate change caused by greenhouse gas emissions from fossil fuels are the major challenges facing mankind today. Biological hydrogen recovery from renewable sources is expected to alleviate human dependence on fossil fuels while mitigating greenhouse gas emissions.¹ Biological hydrogen production relies on functional microbes that convert organic waste into hydrogen through microbial fermentation. Photofermentative bacteria (PFB) can theoretically convert 100% of electrons from a substrate to hydrogen (12 mol H₂ per mol glucose) through utilizing energy from the sun to drive thermodynamically unfavorable reactions.² However, microbial fermentation must consume a significant fraction of the substrate for the production of biomass biocatalysts.³ In addition, hydrogen yields by growing PFB are low since the majority of the electrons from organic substrates are used for biosynthesis.⁴ Therefore, biomass accumulation was identified as a major electron sink during hydrogen production by PFB.⁵

In the conventional dark fermentation, microorganisms obtain energy for cell growth through oxidizing organic substrate into desirable products,⁶ such as volatile fatty acids, hydrogen or methane. This indicates that the degradation of organic substrates, the formation of the products, and the cell growth are coupled together during dark fermentation. However, it must be recognized that PFB do not acquire any energy or reducing equivalents from hydrogen production, while PFB capture energy from sunlight to oxidize organic compounds and generate electron potential needed to drive hydrogen production. The energy harvested from sunlight is also used for biomass synthesis from organic substrate. As a result, cell growth and hydrogen production are not strictly coupled together in photo fermentation,⁷ but compete for the energy and reductant. In a non-growing photoheterotrophic bacterium, metabolic flow was diverted from biosynthesis to mobilize electrons for hydrogen production.⁸ In the stationary growth phase, biomass synthesis was minimal, and the substrate was almost completely converted into hydrogen with highest hydrogen yield of 3.67 mol H₂ per mol acetate.⁹ These non-growing cells shifted their metabolism from the glyoxylate cycle exclusively used by growing cells to the tricarboxylic acid cycle to metabolize acetate,⁸ which enabled cells to more fully oxidize acetate, providing more reducing power to approach the theoretical hydrogen yield. In addition, maintenance of a proton gradient and ATP levels by cyclic photophosphorylation make PFB to remain active in a non-growing state for long periods of time.⁴ Consequently, minimization of biomass production is able to free additional reductant and ATP for hydrogen production through photo fermentation in long term operation. However, most of photo fermentative processes were based on cultivating bacteria in the planktonic state within liquid media, due to the poor flocculation of PFB.^10–12 The PFB were free-floating in the photobioreactor, so the stable suspension of bacterial cells was continuously discharged from photobioreactor with the effluent. In steady state operation, PFB have to keep growing to maintain a constant biomass concentration in photobioreactor. As a result, substrate used for hydrogen production was only account for 16–45% of total substrate consumption, since the majority of substrate was used for biosynthesis.^10–12 Therefore, the poor flocculation of PFB resulting in continuous washout of biomass diverts electron flux away from hydrogen production to continuing biosynthesis.

In previous study, the aggregation of Rhodopseudomonas faecalis RLD-53 induced by L-cysteine was successfully developed in a photobioreactor.¹³ The aggregation mechanism was revealed that through formation of disulfide bonds, L-cysteine promoted production of extracellular polymeric substances (EPS), subsequently changed the cell surface properties and enhanced the adhesion among bacteria cells to form a stable aggregate of PFB. The formation of microbial aggregate is critical to microbial biomass/effluent separation in biological wastewater treatment process.¹⁴ After bioflocculation, microbial biomass responsible for hydrogen production could be retained and accumulated in the photobioreactor to produce hydrogen continuously. The aggregation behavior of microbial cells was subject to influence of substrate concentration, carbon to nitrogen ratio and agitation rate. Therefore, this work investigated the flocculation behavior of R. faecalis RLD-53 under different conditions. EPS and flocculability of PFB were examined for better understanding the role of these conditions in formation of flocculation. Finally, continuous hydrogen production performance by aggregation of R. faecalis RLD-53 was evaluated in a photo fermentative sequencing batch reactor.

2. Experimental

2.1 Bacterium, medium and culture conditions

The photo fermentative bacterium used in this study was Rhodopseudomonas faecalis RLD-53, which was isolated from fresh water pond sediments and had excellent ability for hydrogen production.¹⁵ The culture medium was prepared as described in previous report.¹³ The sodium acetate was used as a sole carbon source, and sodium glutamate was used as nitrogen source. The initial pH of the medium was adjusted to 7.0 using 0.1 mol l⁻¹ NaOH or HCl. The batch culture experiments were carried out in triplicate with 80 ml of the medium in 100 ml sealed reactors and anaerobic conditions were maintained by filling reactors with argon. The reactors with medium were autoclaved at 121 °C for 15 min. R. faecalis RLD-53 in the mid-exponential growth phase was inoculated into reactors at the ratio of 10% (v/v). The light intensity on the outside surface of the reactors was maintained at 150 W m⁻² by incandescent lamps (60 W). The reactors were stirred at 120 rpm at constant temperature of 35 ± 1 °C.

2.2 Photo fermentative sequencing batch reactor

The continuous hydrogen production experiments were conducted in 600 ml sealed glass photo fermentative sequencing batch reactors (PFSBR) with 500 ml working volume. The cylindrical photobioreactor was made of 2 mm glass with the inner diameter of 80 mm and height of 120 mm, respectively. The reactors were flushed 10 min using argon gas with high purity (99.99%) to maintain anaerobic conditions and autoclaved at 121 °C for 15 min. PFSBR has been demonstrated as high biomass retaining process which decouples solid retention time from hydraulic retention time (HRT) and maintains high biomass concentration.¹⁶ The culture conditions were based on the optimum substrate concentration, C/N ratio and agitation rate obtained in batch experiments. During start-up, R. faecalis RLD-53 in the mid-exponential growth phase was inoculated into reactors. Initially, the reactors were operated in batch model for 4 days to accumulated biomass for continuous operation. After that, PFSBR was continuously operated at HRT of 96 h corresponding to 48 h per cycle, including the following four sequential steps: feed (15 min), react (47 h), settle (30 min), and decant (15 min). The feed and decant volume were 250 ml, half working volume of photobioreactor. At each operation condition, the PFSBR was operated over enough time to allow steady-state conditions.

2.3 Flocculability tests

The bacterial cells cultured at different conditions were harvested by centrifugation at 12 000 rpm for 10 min and washed twice with 0.9% NaCl solution. The bacterial cells were resuspended in 0.01 mol l⁻¹ NaCl solution for the flocculability tests. The absorbance of prepared cell suspensions was measured with a spectrophotometer (Shimadzu UV-2550; Shimadzu, Kyoto, Japan) at 650 nm (A₀). Thereafter, the cell suspensions were centrifuged at 1000 rpm for 2 min, and the supernatant optical density was measured again at 650 nm (A_t). Thus, the flocculability of R. faecalis RLD-53 value can be calculated using the following equation:¹⁷

2.4 EPS extraction and analysis

EPS was extracted using cation exchange resin (CER)¹⁸ (Dowex Marathon C, 20–50 mesh, sodium form, Fluka 91973). The bacterial cells were collected through centrifugation at 12 000 rpm, 4 °C for 10 min, and the supernatant was decanted to remove the soluble microbial products in the bulk medium. And then the cell pellets were washed twice using 0.9% NaCl solution. Subsequently, the cells were re-suspended to their original volume with Milli-Q water and transferred to an extraction beaker. The CER was added to the extraction beaker at the ratio of 70 g g⁻¹ cell dry weight, and the mixture was stirred at 600 rpm for 12 h at 4 °C. The samples were centrifuged at 12 000g for 30 min followed by filtration using a 0.45 μm cellulose acetate membrane to remove resin, microorganisms, and residual debris to obtain EPS samples. The chemical compositions of EPS were analyzed following the procedure described in previous report.¹³ The polysaccharide content in EPS was determined by the anthrone method using glucose as a standard. The protein and humic substance in EPS were measured followed the modified Lowry method using bovine serum albumin and humic acid (Fluka Chemical Corp., USA) as the respective standards. The nucleic acid content was measured by the diphenylamine colorimetric method using fish DNA as the standard.

2.5 Analytical method

Light intensity was measured at the surface of reactor with solar power meter TENMARS TM-207 (Tenmars Electronics CO., LTD., Taiwan, China). Biogas was sampled from the head space of the photo bioreactor by using gas-tight glass syringes and hydrogen content was determined by using a gas chromatograph (Agilent 4890D, Agilent Technologies, USA). The gas chromatograph column was Alltech Molesieve 5A 80/100. Argon was used as the carrier gas with a flow rate of 30 ml min⁻¹. Temperatures of the oven, injection, detector, and filament were 35, 120, 120, 140 °C, respectively. Residual substrate in culture broth was determined using a second gas chromatograph (Agilent 7890 A, Agilent Technologies, USA) equipped with a flame ionization detector. The liquor samples were firstly centrifuged at 12 000 rpm for 5 min, and filtered through a 0.2 μm membrane before analysis. The operational temperatures of the injection port, the column and the detector were 220, 190 and 220 °C, respectively. Nitrogen was used as carrier gas at flow rate of 50 ml min⁻¹. The biomass concentration was determined by filtering culture broth through a cellulose acetate membrane filter (0.45 μm pore size, 50 mm in diameter). After that, the filter was rinsed by deionized water to remove salts or non-cellular materials. Each loaded filter was dried at 105 °C until the weight became consistent. The dry weight of blank filter was subtracted from that of the loaded filter to obtain the cell biomass.

3. Results and discussion

3.1 Effect of substrate concentration on the flocculability of R. faecalis RLD-53

Nutrient levels have a significant effect on the production and composition of EPS,¹⁹ while EPS play a critical role in the process of bioaggregate formation. In this study, the sodium acetate at concentrations of 12.5, 25, 50, and 75 mmol l⁻¹ were used in the medium to determine the effect of substrate concentration on EPS production and aggregation of R. faecalis RLD-53 (Fig. 1).

Fig. 1 Effect of substrate concentration on EPS production and aggregation ability: (a) EPS production; (b) aggregation ability.

The total EPS increased with the acetate concentration, and reached the maximum 58.94 mg g⁻¹ at acetate of 50 mmol l⁻¹ (Fig. 1a). The mathematical model²⁰ simulating the kinetic of EPS formation shows that the EPS production rate is proportional to the substrate utilization rate and a higher substrate concentration results in a greater concentration of EPS. The EPS content of the sludge was also found to increase with an increase in food to microorganism ratio in a membrane bioreactors.²¹ EPS were able to physically bridge neighboring cells through altering the properties of the bacterial surface.²² As a result, aggregation was promoted by the production of EPS. Thus, flocculability of R. faecalis RLD-53 increased with the increasing concentration of acetate, and arrived at 36.86% when the optimum concentration was at 50 mmol l⁻¹ (Fig. 1b). In addition, the ratio of proteins to polysaccharides (PN/PS) in EPS has been suggested as a good indicator of the aggregate settle ability and strength.²¹ The PN/PS ratio of the PFB aggregate also reached peak value 0.95 at 50 mmol l⁻¹ of acetate, indicating that the aggregate of PFB cultivated at 50 mmol l⁻¹ of acetate probably have better settle ability and mechanical strength. However, with the further increase to 75 mmol l⁻¹, EPS production of R. faecalis RLD-53 decreased to 42.88 mg g⁻¹, which resulted in the decrease of flocculability to 23.45%. This may be due to the substrate concentrations exceed the inhibitory threshold concentrations. This phenomena was also reported by Chen²³ that the aggregate of anammox turn into flocculent sludge with the excessive concentration of substrate.

Fig. 2 Effect of carbon to nitrogen ratio on EPS production and aggregation ability: (a) EPS production; (b) aggregation ability.

3.2 Effect of C/N ratio on the flocculability of R. faecalis RLD-53

The C/N ratio plays an important role in a biological process, which is a key indicator for nutrient balance of culture medium.²⁴ The C/N ratio affected the production and constituents of EPS, and subsequently the cell surface properties. The effect of C/N ratio was investigated by sequentially increasing the glutamate concentration from 0, 5, 10 to 20 mmol l⁻¹ in each test while maintaining the acetate concentration constant at 50 mmol l⁻¹, corresponding to C/N ratio at 10 : 0, 10 : 1, 10 : 2 and 10 : 4, respectively.

The EPS production increased with the decrement of C/N ratio, and arrived at maximum of 68.14 mg g⁻¹ at C/N ratio of 10 : 4 (Fig. 2a). The polysaccharides (19.58 mg g⁻¹) were the dominant component in EPS at C/N ratio of 10 : 0, while the protein was only 6.97 mg g⁻¹. As the C/N ratio decreased, more extracellular protein were produced. Durmaz and Sanin²⁵ also found EPS in activated sludge to be rich in proteins but low in polysaccharides at a carbon to nitrogen ratio of 5, but as the carbon to nitrogen ratio increased to 40, the amount of protein decreased sharply whereas the amount of polysaccharides increased. In this study, the protein content in EPS increased from 6.97 to 26.71 mg g⁻¹ with the decrease of C/N ration from 10 : 0 to 10 : 4, while polysaccharide slightly increased from 19.58 to 26.93 mg g⁻¹.

Proteins have been demonstrated to play a crucial role in the aggregation of bacterial cell. Amino acid analysis and amino acid sequencing of protein in EPS suggested the protein was a lectinlike protein and binding site inhibition studies showed that the protein had lectinlike activity.²⁶ In this study, the flocculability of R. faecalis RLD-53 was substantially influenced by the C/N ratio (Fig. 2b) with the changes of EPS covered on cell surface. Flocculability increased with the decrement of C/N ratio, and reached maximum 40.79% at C/N ratio of 10 : 2. However, the further increase of C/N ratio resulted in a sharp decrease of flocculability to 33.45%. Ye²⁷ also reported that flocculability increased with the decrease of C/N ratio from 100 to 20, while the turbidity, effluent suspended solids content and sludge volume index increased greatly at the C/N ratio of 4, indicating seriously deteriorated flocculation at this ratio. Therefore, the results indicated that a proper concentration of EPS at C/N ratio of 10 : 2 was beneficial to aggregation of R. faecalis RLD-53, but with the excessive production of EPS at C/N of 10 : 4, flocculability deteriorated.

3.3 Effect of agitation rate on the flocculability of R. faecalis RLD-53

Agitation rate is a key parameter affecting the mass transfer in biological systems,²⁸ and also plays an important role in aggregation formation through initiating bacterium to bacterium contact.²⁹ In this study, EPS production and aggregation of PFB were examined at different agitation rates of 0, 40, 80, 120 and 160 rpm to investigate the effect of agitation on the aggregation behavior of strain RLD-53. In all tests, acetate concentration was 50 mmol l⁻¹ with C/N ratio of 10 : 2.

The agitation rate had a substantially effect on the flocculation behavior of R. faecalis RLD-53 (Fig. 3). At agitation rate of 0 rpm, collisions among microbial cell were dominated by Brownian motion, but Brownian motion is only appreciable when particles size in submicron.³⁰ The cell radius of R. faecalis RLD-53 was 2.38 μm, and aggregate grow larger, so that Brownian motion was weak enough to be neglected. As a result, bacterial cells were in stable suspension with flocculability only 24.31%. The increment of agitation rate could induce more effective collision among microbial cells, which facilitated the aggregation of PFB. As a result, the flocculability of R. faecalis RLD-53 increased with the increase of agitation rate, and arrived at maximum of 47.67% at 80 rpm. Hydrodynamic shear force generated from agitation has been demonstrated to promote the secretion of EPS,^31,32 which was favorable to formation of bioaggregate in wastewater treatment system. Consequently, total EPS content increased with the increment of agitation rate and reached maximum of 44.71 mg g⁻¹ at 80 rpm. However, intensive hydrodynamic shear force generated from high agitation rate overwhelmed the adhesion among microbial cells and resulted in disruption of PFB aggregate, which gave the minimum flocculability at 16.17%. In addition, intensive hydrodynamic shear also caused a considerable amount of EPS released from the bacterial surface. Menniti³³ reported that biomass grown in a high shear environment had lower floc-associated EPS production than biomass grown in a lower shear environment. In this study, the total EPS of R. faecalis RLD-53 cultivated at 160 rpm was only 26.12 mg g⁻¹. From the above results, the suitable agitation was found to be 80 rpm, since it gave stable bioaggregates.

Fig. 3 Effect of agitation rate on EPS production and aggregation ability: (a) EPS production; (b) aggregation ability.

3.4 Continuous hydrogen production by aggregate of R. faecalis RLD-53

Based on the above experimental results, the optimum flocculability was obtained at substrate concentration of 50 mmol l⁻¹, C/N ratio of 10 : 2 and agitation rate of 80 rpm. The continuous hydrogen production performance by the strain RLD-53 with the excellent flocculability in PFSBR was evaluated. The control was free cell culture, which was strain RLD-53 cultivated at same conditions without addition of L-cysteine.

Fig. 4a showed R. faecalis RLD-53 in free cell culture, indicating that biomass was in stable suspension, and there was almost no difference after 30 minutes settling (Fig. 4b). As a result, about half of biomass was washed away with effluent at each decant stage (Fig. 5c). During the whole operation, 2.86 g of biomass was discharged from photobioreactor while only 0.92 g was retained in photobioreactor at the end of operation. Moreover, the washout of biomass was supposed to further increase with increasing running time. The low concentration of hydrogen producer not only resulted in poor hydrogen production rate of 374 ml l⁻¹ d⁻¹, but also caused insufficient substrate removal with 12.08 mmol l⁻¹ of acetate in the effluent. After fresh medium was introduced, acetate was continuously utilized for biomass synthesis to supplement the biomass washout. Thus, the lower hydrogen yield of 1.80 mol H₂ per mol acetate was detected, and this suggested the electrons from organic substrate for PFB cell biosynthesis was more than that for hydrogen production. In addition, the presence of biomass and residual substrate with high concentration in the effluent increased the effluent turbidity, meanwhile led to increases in pollutants like chemical oxygen demand, total nitrogen, and total phosphate, causing poor effluent water quality. Thus, the effluent from photobioreactor required further treatment before discharge, which will greatly increase the operating costs.

Fig. 4 Photographs of R. faecalis RLD-53 cultivated at different conditions: free cell culture at react stage (a) and settle stage (b); bioaggregate of PFB at react stage (c) and settle stage (d).

Fig. 5 Continuous hydrogen production by bioaggregate and free cell of RLD-53 in photo fermentative sequencing batch reactors: (a) hydrogen yield; (b) hydrogen production rate; (c) biomass concentration.

In contrast, the aggregate of PFB fluidized in photobioreactor in the react stage where the medium was agitated at 80 rpm (Fig. 4c), which enabled each bacteria in the bioaggregate could receive light energy and convert substrate to hydrogen. With the excellent flocculability, the aggregate formed by PFB was allowed to settle to the bottom of photobioreactor during the settle stage (Fig. 4d). Only a negligible proportion of biomass was discharged with the effluent as free cells at decant stage (Fig. 5c), which indicated that the most of PFB accumulated in a photobioreactor to produce hydrogen continuously. As a result, the biomass gradually accumulated in the photobioreactor, and achieved 1.55 g l⁻¹ (Fig. 5b). In addition, during the entire operation, only 0.55 g of biomass was washed away from photobioreactor. The results indicated that the biomass synthesis was greatly reduced compared with that of free cell culture. Accordingly, hydrogen production rate of bioaggregate gradually increased with the accumulation of biomass and reached maximum value of 706 ml l⁻¹ d⁻¹ (Fig. 5b).

Using solar energy to drive thermodynamically unfavorable reactions, PFB could completely convert organic waste into hydrogen. And the effluent from photobioreactor should also contain fewer by-products as the organic feedstock are fully converted into H₂ and CO₂.³⁴ However, continuous biomass washout from photobioreactor due to the poor flocculation of PFB not only deteriorates the effluent quality, but also diverts electron flux away from hydrogen production to continuing biosynthesis. Up to the present, cell entrapment³⁵ and biofilm formation on solid carriers³⁶ have been applied into photobioreactor in order to enhance the biomass retention capacity with consequent increases in reaction rates and productivity. Entrapment of PFB in polymeric matrices allows high cell density under low HRT and creates a local anaerobic environment for hydrogen fermentation.³⁷ However, cell entrapment is not feasible in long term operation, because mass transfer limitations and light penetration limitations arise from dense packing of cell entrapment³⁸ and structural damage caused by biogas production. In addition, the porous glass, expanded clay and granule activated carbon, reactor surface have been used for the photo fermentative bacteria immobilization.^39,40 Due to high density, these solid carriers deposited on the reactor bottom, and resulted in serious shading effect on each other, which greatly decreased the hydrogen production performance. In this study, stable floc of PFB induced by L-cysteine was developed (Fig. 4c and d) under the optimum culture conditions with substrate concentration of 50 mmol l⁻¹, C/N ratio of 10 : 2 and agitation rate of 80 rpm. During the react stage, the proper agitation kept the aggregate of PFB fluidizing in photobioreactor, which indicated that each cell in the bioaggregate could receive light energy and convert substrate to hydrogen. During settle stage of reactor operation, aggregate of PFB settled to the bottom of the reactor, and accumulated in the photobioreactor (Fig. 5). As a result, a negligible proportion of biomass was discharged with the effluent as free cells, and biomass synthesis was greatly reduced. Consequently, a considerable fraction of the electrons from organic substrate was diverted away from biomass synthesis into hydrogen production. Thus, maximum hydrogen production rate 706 ml l⁻¹ d⁻¹ were obtained with hydrogen yield of 2.57 mol H₂ per mol acetate. Therefore, the aggregation of R. faecalis RLD-53 is a promising strategy to enhance hydrogen recovery from organic substrate through circumventing electron sink to biomass synthesis.

In addition, hydrogen production performance of photobioreactor was demonstrated to be significantly influenced by the hydraulic retention time, organic loading rate and the light intensity. Thus, a further investigation on the operational condition of PFSBR should be carried out to maximize hydrogen production from organic waste.

4. Conclusions

In this study, the influences of substrate conditions, C/N ratio and agitation rate on the aggregation behavior of R. faecalis RLD-53 were investigated. The results indicated the substrate concentration, C/N ratio and agitation rate determined the aggregation ability of PFB. The maximum flocculability of 47.67% was obtained at following optimal conditions, substrate concentration of 50 mmol l⁻¹, C/N ratio of 10 : 2 and agitation rate of 80 rpm. As a result, biomass washout was reduced from 2.86 g for free cell culture to 0.55 g for bioaggregate of PFB in whole continuous operation, which minimized the substrate consumption for biomass synthesis. Consequently, bioaggregate of PFB reached a higher steady-state hydrogen production rate of 706 ml l⁻¹ d⁻¹ and hydrogen yield of 2.57 mol H₂ per mol acetate, respectively. Therefore, aggregation of R. faecalis RLD-53 is a promising strategy to maximize reductant flow from organic substrate into hydrogen production through circumventing electron sink to biomass synthesis.

Acknowledgements

The authors thank the Natural Science Foundation of China (51478139), Harbin Innovation Talents Funding of Science and Technology (2014RFQXJ084), Fundamental Research Funds for the Central Universities (HIT.BRETIII.201418), State Key Laboratory of Urban Water Resource and Environment (Harbin Institute of Technology) (2015TS02) and the Shanghai Tongji Gao Tingyao Environmental Science & Technology Development Foundation for supporting this study.

Notes and references

1. I. Gelfand, R. Sahajpal, X. Zhang, R. C. Izaurralde, K. L. Gross and G. P. Robertson, Nature, 2013, 493, 514–517.
2. L. B. Brentner, J. Peccia and J. B. Zimmerman, Environ. Sci. Technol., 2010, 44, 2243–2254.
3. W. D. Huang and Y. H. Percival Zhang, Energy Environ. Sci., 2011, 4, 784–792.
4. J. B. McKinlay and C. S. Harwood, Curr. Opin. Biotechnol., 2010, 21, 244–251.
5. W. S. Kontur, D. R. Noguera and T. J. Donohue, Curr. Opin. Biotechnol., 2012, 23, 382–389.
6. D. J. Batstone, J. Keller, I. Angelidaki, S. V. Kalyuzhnyi, S. G. Pavlostathis, A. Rozzi, W. T. M. Sanders, H. Siegrist and V. A. Vavilin, Water Sci. Technol., 2002, 45, 65–73.
7. M. R. Melnicki, L. Bianchi, R. de Philippis and A. Melis, Int. J. Hydrogen Energy, 2008, 33, 6525–6534.
8. J. B. McKinlay, Y. Oda, M. Rühl, A. L. Posto, U. Sauer and C. S. Harwood, J. Biol. Chem., 2014, 289, 1960–1970.
9. G. J. Xie, B. F. Liu, D. F. Xing, J. Ding, J. Nan, H. Y. Ren, W. Q. Guo and N. Q. Ren, Int. J. Hydrogen Energy, 2012, 37, 13718–13724.
10. J. J. Gilbert, S. Ray and D. Das, Int. J. Hydrogen Energy, 2011, 36, 3434–3441.
11. A. A. Tsygankov, A. S. Fedorov, T. V. Laurinavichene, I. N. Gogotov, K. K. Rao and D. O. Hall, Appl. Microbiol. Biotechnol., 1998, 49, 102–107.
12. E. Boran, E. Özgür, J. van der Burg, M. Yücel, U. Gündüz and I. Eroglu, J. Cleaner Prod., 2010, 18(suppl. 1), S29–S35.
13. G. J. Xie, B. F. Liu, D. F. Xing, J. Nan, J. Ding and N. Q. Ren, Biotechnol. Biofuels, 2013, 6, 64.
14. X. M. Liu, G. P. Sheng, H. W. Luo, F. Zhang, S. J. Yuan, J. Xu, R. J. Zeng, J. G. Wu and H. Q. Yu, Environ. Sci. Technol., 2010, 44, 4355–4360.
15. N. Q. Ren, B. F. Liu, J. Ding and G. J. Xie, Bioresour. Technol., 2009, 100, 484–487.
16. G. J. Xie, B. F. Liu, W. Q. Guo, J. Ding, D. F. Xing, J. Nan, H. Y. Ren and N. Q. Ren, Int. J. Hydrogen Energy, 2012, 37, 13689–13695.
17. X. M. Liu, G. P. Sheng and H. Q. Yu, Environ. Sci. Technol., 2007, 41, 4620–4625.
18. K. Y. Chan, L. C. Xu and H. H. P. Fang, Environ. Sci. Technol., 2002, 36, 1720–1727.
19. G. P. Sheng, H. Q. Yu and X. Y. Li, Biotechnol. Adv., 2010, 28, 882–894.
20. B. J. Ni and H. Q. Yu, AIChE J., 2010, 56, 546–559.
21. B. S. Xing, T. Y. Qin, S. X. Chen, J. Zhang, L. X. Guo and R. C. Jin, Bioresour. Technol., 2013, 149, 310–317.
22. Y. Q. Liu, Y. Liu and J. H. Tay, Appl. Microbiol. Biotechnol., 2004, 65, 143–148.
23. T. T. Chen, P. Zheng, L. D. Shen, C. J. Tang and S. Ding, Biotechnol. Bioprocess Eng., 2012, 17, 1093–1102.
24. R. Li, Z. Zhi and H. Wang, J. Renewable Sustainable Energy, 2014, 6, 033139.
25. B. Durmaz and F. Sanin, Water Sci. Technol., 2001, 44, 221–229.
26. M. J. Higgins and J. T. Novak, J. Environ. Eng., 1997, 123, 479–485.
27. F. Ye, Y. Ye and Y. Li, J. Hazard. Mater., 2011, 188, 37–43.
28. S. Khanna, A. Goyal and V. S. Moholkar, Prep. Biochem. Biotechnol., 2013, 43, 828–847.
29. L. Zhu, H. Y. Qi, M. L. Lv, Y. Kong, Y. W. Yu and X. Y. Xu, Bioresour. Technol., 2012, 124, 455–459.
30. K. A. Kusters, J. G. Wijers and D. Thoenes, Chem. Eng. Sci., 1997, 52, 107–121.
31. J. H. Tay, Q. S. Liu and Y. Liu, Appl. Microbiol. Biotechnol., 2001, 57, 227–233.
32. J. Wu, H. M. Zhou, H. Z. Li, P. C. Zhang and J. Jiang, Water Res., 2009, 43, 3029–3036.
33. A. Menniti, S. Kang, M. Elimelech and E. Morgenroth, Water Res., 2009, 43, 4305–4315.
34. G. J. Xie, B. F. Liu, H. Y. Ren, D. F. Xing, J. Nan and N. Q. Ren, GCB Bioenergy, 2014, 6, 621–628.
35. B. F. Liu, G. J. Xie, W. Q. Guo, J. Ding and N. Q. Ren, Nat. Resour., 2011, 2, 1.
36. G. J. Xie, B. F. Liu, J. Ding, H. Y. Ren, D. F. Xing and N. Q. Ren, Int. J. Hydrogen Energy, 2011, 36, 13991–13996.
37. G. J. Xie, B. F. Liu, J. Ding, D. F. Xing, H. Y. Ren, W. Q. Guo and N. Q. Ren, Biomass Bioenergy, 2012, 44, 122–129.
38. X. Tian, Q. Liao, X. Zhu, Y. Z. Wang, P. Zhang, J. Li and H. Wang, Bioresour. Technol., 2010, 101, 977–983.
39. A. A. Tsygankov, Y. Hirata, M. Miyake, Y. Asada and J. Miyake, J. Ferment. Bioeng., 1994, 77, 575–578.
40. C. Y. Chen and J. S. Chang, Process Biochem., 2006, 41, 2041–2049.

---