Hydrogen Production and Consumption of Organic Acids by a Phototrophic Microbial Consortium

Universidade de São Paulo Biblioteca Digital da Produção Intelectual - BDPI

Departamento de Hidráulica e Saneamento - EESC/SHS Artigos e Materiais de Revistas Científicas - EESC/SHS

2012 Hydrogen production and consumption of organic acids by a phototrophic microbial consortium

INTERNATIONAL JOURNAL OF HYDROGEN ENERGY, OXFORD, v. 37, n. 16, supl. 1, Part 3, pp. 11691-11700, AUG, 2012 http://www.producao.usp.br/handle/BDPI/37914

Downloaded from: Biblioteca Digital da Produção Intelectual - BDPI, Universidade de São Paulo international journal of hydrogen energy 37 (2012) 11691e11700

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Hydrogen production and consumption of organic acids by a phototrophic microbial consortium

Carolina Zampol Lazaro*, Daniele Vital Vich, Julia Sumiko Hirasawa, Maria Bernadete Amaˆncio Varesche

Department of Hydraulics and Sanitation, School of Engineering of Saõ Carlos, University of Saõ Paulo, Av. Trabalhador Saõ-carlense, 400, 13566-590 Saõ Carlos, SP, Brazil article info abstract

Article history: Alternative fuel sources have been extensively studied. Hydrogen gas has gained attention Received 1 February 2012 because its combustion releases only water, and it can be produced by microorganisms Received in revised form using organic acids as substrates. The aim of this study was to enrich a microbial 14 May 2012 consortium of photosynthetic purple non-sulfur bacteria from an Upﬂow Anaerobic Sludge Accepted 17 May 2012 Blanket reactor (UASB) using malate as carbon source. After the enrichment phase, other Available online 19 June 2012 carbon sources were tested, such as acetate (30 mmol l 1), butyrate (17 mmol l 1), citrate (11 mmol l 1), lactate (23 mmol l 1) and malate (14.5 mmol l 1). The reactors were incu- Keywords: bated at 30 C under constant illumination by 3 ﬂuorescent lamps (81 mmol m 2 s 1). The 1 Mixed culture cumulative hydrogen production was 7.8, 9.0, 7.9, 5.6 and 13.9 mmol H2 l culture for Rhodobacter acetate, butyrate, citrate, lactate and malate, respectively. The maximum hydrogen yield 1 Sulfurospirillum was 0.6, 1.4, 0.7, 0.5 and 0.9 mmol H2 mmol substrate for acetate, butyrate, citrate, lactate Organic acids and malate, respectively. The consumption of substrates was 43% for acetate, 37% for Anaerobic batch reactors butyrate, 100% for citrate, 49% for lactate and 100% for malate. Approximately 26% of the clones obtained from the Phototrophic Hydrogen-Producing Bacterial Consortium (PHPBC) were similar to Rhodobacter, Rhodospirillum and Rhodopseudomonas, which have been widely cited in studies of photobiological hydrogen production. Clones similar to the genus Sul- furospirillum (29% of the total) were also found in the microbial consortium. Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

1. Introduction Phototrophic bacteria are highlighted in the literature as promising microbial systems for the biological production of Rapid worldwide industrialization and urbanization as well as hydrogen because of their ability to consume organic the increase in fossil fuel consumption have led to some substrates present in wastes, indicating the potential for concerns regarding environment pollution. Thus, attention combining wastewater treatment and energy generation. has been given to alternative sources of energy that produce Simple organic molecules, such as organic acids generated hydrogen as a clean energy carrier. Biological hydrogen from the anaerobic digestion of organic matter, can be used by production processes are known to be more environmental phototrophic bacteria as substrates for H2 production. The friendly and consume less energy compared to thermochem- maximum hydrogen yield from a particular substrate, ical and electrochemical processes because biological assuming complete conversion of the substrate to H2 and CO2, processes take place at ambient temperature and pressure [1]. can be calculated using Equation (1) [2,3].

* Corresponding author.Tel.:þ55 1633738357; fax: þ55 1633739550. E-mail addresses: [email protected] (C.Z. Lazaro), [email protected] (M.B.A. Varesche). 0360-3199/$ e see front matter Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2012.05.088 11692 international journal of hydrogen energy 37 (2012) 11691e11700

þ / þ þ CaHbOc (2a c)H2O (2a c 0.5b)H2 aCO2 (1) According to Li and Fang [6], the development of photo- fermentative hydrogen production is in its initial phase. The substrate conversion efficiency, which is a measure of There are some important aspects that need to be developed how much of the substrate has been utilized for hydrogen by further studies. One of these aspects is the use of mixed production, is another useful parameter. The efficiency is cultures from environmental samples for treating complex determined by calculating the ratio of moles of hydrogen that substrates. have actually been produced to moles of hydrogen expected Thus, the aim of this study was to evaluate the hydrogen from stoichiometric conversion of the substrate [4]. production from organic acids using an enriched phototrophic Most studies of photobiological hydrogen production have microbial consortium as the inoculum. been conducted using pure cultures, and very little informa- tion has been reported regarding photo-fermentative hydrogen production by mixed cultures of phototrophic 2. Materials and methods bacteria [5,6]. The microorganisms most commonly used for the biological production of hydrogen are Rhodobacter sphaer- 2.1. Inoculating and enriching the microbial consortium oides [4,7,8], Rhodobacter capsulatus [9] and Rhodopseudomonas palustris [10e12]. The Phototrophic Hydrogen-Producing Bacterial Consortium Eroglu et al. [13] studied the photo-fermentative hydrogen (PHPBC) was obtained from anaerobic granular sludge from an production from olive mill wastewater (OMW) by R. sphaeroides Upflow Anaerobic Sludge Blanket reactor (UASB) used to treat O.U.001 under two different regimes of illumination, contin- pig slaughterhouse wastewater (Jaboticabal, Brazil). Firstly, uous and 12 h lighte12 h dark. Biomass submitted to contin- the sludge was triturated to break down the granules. Then, uous illumination reached a stationary phase of about 0.55 g 100 ml of the triturated sludge was transferred to a glass bottle dry cell weight l 1 culture at 44 h after inoculation, with (Duran ) with 101 mm of diameter and total volume of 1 l a maximum H2 productivity (based on biomass) of 2.37 ml containing 0.4 l of RCVB culture medium [18]. Helium gas (99%) 1 1 e H2 g h . Initially, under 12 h light 12 h dark cycles, the was purged in the headspace to ensure the anaerobic condi- biomass showed a slow growth rate, which persisted for about tions. The culture medium was supplemented with malate 36 h. After this time, the culture showed an enhancement in (30 mmol l 1) and sodium glutamate (4 mmol l 1). The initial the growth rate, reaching at stationary phase 0.5 g dry cell pH was adjusted to 6.8e7.0 with NaOH (1 N). The bottle was weight l 1 culture at about 70 h after inoculation, with incubated at 30 C in a growth chamber constantly illumi- 1 1 maximum productivity of 4.4 ml H2 g h . The calculated nated by 3 fluorescent lamps (20 W). When a layer of purple yield was the same for the two conditions, with 250 mmol biomass appeared, it was transferred to another glass bottle 1 H2 l OMW. Seifert et al. [14] also investigated the growth of R. using a platinum loop and cultivated under the same condi- sphaeroides O.U.001 on Biebl and Pfenning media using dairy tions described above. This process of transferring the purple wastewater instead of malic acid in its composition. They biomass to another glass bottle with fresh culture medium evaluated the influence of illumination, pH, dairy wastewater was performed about five times to primarily harvest the and inoculum concentrations. The highest hydrogen yield purple non-sulfur photosynthetic bacteria. 1 (3.2 l H2 l ) was reached with light intensities between 9 and 13 klux, 0.086 g dry weight l 1, dairy wastewater at a concen- 2.2. Operation of the batch reactors tration of 40% and pH controlled close to 7. The authors concluded that the amount of hydrogen produced from dairy Microbial hydrogen production experiments were conducted wastewater in these conditions was 30% better than that in closed glass bottles (Duran ) with 86 mm of diameter and 1 observed with standard samples of malic acid (2.3 l H2 l ). total volume capacity of 0.5 ml containing 0.35 l culture Hydrogen production by different Rhodobacter sp. was also medium [19]. Argon gas (99.99%) was purged to ensure the investigated by Kapdan et al. [15] using acid hydrolyzed wheat anaerobic conditions. The substrates were acetate starch by photo-fermentation. The highest specific hydrogen (30 mmol l 1), butyrate (17 mmol l 1), citrate (11 mmol l 1), 1 1 1 1 production rate (46 ml H2 g biomass h ) and the yield lactate (23 mmol l ) or malate (14.5 mmol l ). Sodium 1 1 (1.23 mol H2 mol glucose) were obtained with the R. sphaer- glutamate (4 mmol l ) was used as the nitrogen source. The oides RV culture. concentrations of carbon and nitrogen sources were set to Avcioglu et al. [16] used the dark fermentation effluent of maintain a balanced carbon and nitrogen ratio (C/N ratio molasses as substrate for hydrogen production by two strains equal to an average of 16). The bottles were kept in of R. capsulatus. The system used was an outdoor panel a controlled-temperature incubator at 30 C illuminated photobioreactor. The maximum hydrogen yield obtained was constantly by 3 fluorescent lamps (TLT 20W/75RS, extra day 78% (of the theoretical maximum). Hydrogen productivity of light, Philips). The light intensity at the surface of the bottle photosynthetic bacteria on fermented effluent of potato was approximately 81 mmol m 2 s 1. The photosynthetically steam peels hydrolysate was searched by Afsar et al. [17] active radiation measurement was performed with an LI-193 using strains of R. capsulatus, R. sphaeroides and R. palustris. spherical quantum sensor coupled with datalogger LI-1400 In this study, the authors conclude that using defined co- (LI-COR ). Before each assay, the biomass was centrifuged cultures might be advantageous for overcoming the difficul- (10,000 rpm at 4 C for 5 min) and was washed with fresh ties of varying feedstock properties and may also be benefi- medium. The culture medium [19] used for the hydrogen cial for obtaining a more stable operation for continuous production experiments contained the following nutrients per cultures. liter of deionized water: 0.5 g KH2PO4; 0.4 g MgSO4 7H2O; 0.4 g international journal of hydrogen energy 37 (2012) 11691e11700 11693

NaCl; 50 mg CaCl2 2H2O; 3.9 mg Fe(III) citrate; 0.3 mg H3BO3; selected, incubated in 5 ml of liquid Luria Bertani media (LB) 0.03 mg Na2MoO4 2H2O; 0.1 mg ZnSO4 7H2O; 0.2 mg CoCl2 (37 C, overnight) and submitted to PCR using the primers 0 0 2H2O; 0.01 mg CuCl2 2H2O, 0.03 mg MnCl2 4H2O; 0.02 mg NiCl2 M13F (5 -CGC CAG GGT TTT CCC AGT CAC GAC-3 ) and M13R 0 0 6H2O and 0.04 mg vitamin B12. The initial pH was adjusted to (5 -TTT CAC ACA GGA AAC AGC TAT GAC-3 ) [23]. The PCR 6.8e7.0 with NaOH (1 N). products were purified with Illustra GFX (GE Healthcare) kit. Nucleotide sequencing was performed on an automated ABI Prism 310 sequencer (Dye terminator cycle sequencing kit, 2.3. Analytical methods Applied Biosystems) according to the manufacturer’s instructions, using the M13F primer. The hydrogen content in the biogas was determined by The nucleotide sequences analysis was processed accord- analyzing a gas samples (250 ml) from each reactor using a gas ing to the procedure described by Oliveira et al. [24]. The chromatograph (GC 2010, Shimadzu, Tokyo, Japan) equipped phylogenetic tree was constructed using PHYLIP 3.6 and was with a Carboxen 1010 PLOT column (30 m, 0.53 mm), a thermal calculated using the neighbor-joining algorithm [25]. Boot- conductivity detector (TCD) and argon as a carrier gas. The strap re-sampling analysis for 100 replicates was performed to temperatures of the injector and the detector were 130 C and estimate the confidence of tree topologies. 230 C, respectively. The column temperature program was as The nucleotide sequences determined in this study and follows: 40 C for 2 min, heated to 60 C at the rate of included in the phylogenetic tree have been deposited in 5 Cmin 1, followed by an increase to 95 C at the rate of GenBank under the accession numbers JQ437239eJQ437255. 25 Cmin1, and then increasing at 5 Cmin1 to a final temperature of 200 C. 2.4.3. Modeling hydrogen production using the Gompertz Organic acids concentrations were measured using a high equation performance liquid chromatograph (HPLC) (Shimadzu) The cumulative hydrogen production (H ) data were fitted to equipped with an Aminex HPX-87H column (300 mm, 7.8 mm; a modified Gompertz equation [26], which is a suitable model Bio-Rad) with column oven (CTO-20A) at 64 C, UV-diode array for describing the progressive accumulation of hydrogen in detector (SPD-M10A VP), system controller (SCL-10A VP) and a batch experiment (Equation (3)): pump (LC-10AD VP). The mobile phase was H2SO4 solution 1 (0.01 N) at 0.8 ml min . The cellular growth was determined R $e H ¼ P$exp exp m ðl tÞþ1 (3) spectrophotometrically based on optical density at 660 nm P

(OD660nm) (DR/4000 Spectrophotometer-HACH) and the where P is the hydrogen production potential (mmol H l 1 biomass dry weight was measured as volatile suspended 2 1 culture), Rm is the maximum H2 production rate (mmol H2 l solids (VSS) [20]. Standard-growth curve experiment was culture d 1), l is lag-phase time (d) and e is 2.718281828. This carried out in order to establish the ratio between cell density curve-ﬁtting was accomplished using Statistica 8.0 software. and biomass dry weight. In this experiment, there were The experiments were carried out in duplicates. collected samples for the optical cell density and VSS analysis. After that the data were plotted and it was found that an optical cell density of 1.0 corresponded to 0.92 g dry weight l 1 culture (Equation (2)). 3. Results and discussion

1 ¼ þ Many studies have been done using pure cultures of photo- VSS (g l ) 0.9204 OD660 nm 0.0019 (2) trophic purple bacteria to the photo-fermentative hydrogen production. However, the hydrogen production performed by 2.4. Molecular biology microbial consortia must be studied as a way to evaluate the economic viability of the process. Hence, this study evaluated 2.4.1. Nucleic acid extraction the photo-fermentative hydrogen production by microbial The analysis of the microbial diversity was performed using consortium from organic acids. sample collected at the end of the enrichment phase. DNA extraction was performed according to the procedure 3.1. Enrichment of Phototrophic Hydrogen-Producing described by Griffiths et al. [21], modified by the use of glass Bacterial Consortium beads and a phenol:chloroform:buffer mixture (1:1:1 v/v). A segment of the 16S rRNA gene was amplified by PCR using the The Phototrophic Hydrogen-Producing Bacterial Consortium eubacterial primers 27F (50-AGA GTT TGA TCC TGG CTC AG-30) (PHPBC) obtained was composed mainly of Gram-negative and 1100R (50-AGG GTT GCG CTC GTT G-30) [22]. rods, but Gram-positive rods were found as well. Approxi- mately 33% of the clones were similar to bacteria belonging to 2.4.2. Cloning and determination of the 16S rRNA sequence Alphaproteobacteria, including Rhodobacter, Rhodospirillum The PCR products were cloned using the pGEM Easy Vector and Rhodopseudomonas [27]. These bacterial genera have been System (Promega) according to the manufacturer’s specifica- widely reported in studies focusing on hydrogen production tions. For this purpose, 2 ml of PCR product was used in the by phototrophic purple non-sulfur bacteria [6]. Clones similar ligation with pGEM vector and subsequent transformation in to the genus Sulfurospirillum were also found in the microbial competent cells. 100 ml of the transformated culture was consortium. Sulfurospirillum are able to consume acetate, plated onto LB/Ampicilin/IPTG/X-Gal plates and incubated butyrate, citrate, lactate and malate, but they cannot produce overnight at 37 C. After this period, 50 clones were randomly hydrogen gas. Moreover, these microorganisms can use 11694 international journal of hydrogen energy 37 (2012) 11691e11700

hydrogen gas as an electron donor in the presence of acetate 81% of species were closely related to R. capsulatus (99% [28]. similarity), while two unidentiﬁed species were related to the The phylogenetic tree was constructed using 42 sequences Bacillus/Clostridium. These microorganisms most likely prop- obtained by phylogenetic analysis of 16S rRNA gene fragments agated in the microbial consortium due to their metabolic (Fig. 1). These 42 sequences were grouped into 17 operational versatility and their ability to sporulate, although the envi- taxonomic units (OTUs), with 97% similarity (Fig. 2). ronmental conditions imposed in this study were not suitable OTU_1 is comprised of a clone similar to Acholeplasma sp. for the enrichment of these microorganisms. (EU517562 e 92%), a microorganism of the class Mollicutes OTU_7 and OTU_8 contained 2 clones each, which were (phylum Firmicutes), which contains bacteria that have no cell similar to uncultured bacteria (AY945866 e 94%). OTU_9 con- wall. They are not related to the production of hydrogen and tained 3 clones similar to Propionicimonas paludica (EF515420 e are described in the literature as pathogens of plants and 89.5%), which is capable of fermenting glucose to produce animals [29]. The source of the inoculum used in this study acetic, lactic, succinic and propionic acids and CO2. Akasaka may explain the presence of this microorganism in the et al. [33] isolated two strains under anaerobic conditions that microbial consortium. were able to utilize several carbon sources (arabionose, xylose, Clones similar to uncultured anaerobic bacteria belonging fructose, glucose and sucrose) and also had the ability to grow to the class Clostridia (phylum Firmicutes) were grouped into well on pyruvate and lactate and less well in malate, fumarate two different OTUs. One clone was placed in OTU_2 (AY953192 and succinate. e 100%), and other 5 clones were grouped in OTU_16 OTU_13 contained 3 clones similar to uncultured Magne- (EU864476 e 100%). Clostridium species are Gram-positive tospirillum sp. (EF176784 e 86%), which is a genus belonging to spore-forming bacteria that can produce hydrogen by the the class Alphaproteobacteria. They are unable to grow fermentative metabolism of carbohydrates. They are photosynthetically but can use organic acids as a nutrient frequently reported in the literature as hydrogen-producing source [34], which explains their presence in the phototrophic bacteria that use organic molecules (like sucrose and microbial consortium; however, these microorganisms have glucose) as carbon sources and produce H2,CO2, organic acids not been described in the literature as hydrogen producers. and alcohols as byproducts [30e32]. Zhang et al. [5], performed OTU_14 contained 3 clones similar to Propionivibrio e a DNA cloning analysis of a phototrophic sludge producing H2 (AY928207 90%). Tanaka et al. [35] isolated the strain Cre- from acidiﬁed wastewater and reported that approximately Mal1, which they described as a member of Propionivibrio

Sulfurospirillum deleyanum (AB368775) OTU_17 Epsilonproteobacteria Sulfurospirillum cavolei (AB246781) Sulfurospirillum sp. (AY780560) OTU_15 OTU_14 Betaproteobacteria Propionivibrio sp. (AY928207) OTU_6 OTU_5 Rhodospirillum photometricum (D30777) Rhodopseudomonas palustris (AB017261) OTU_4 Rhodoplanes cryptolactis (AB087718) OTU_3 Magnetospirillum sp. (EF176784) Alphaproteobacteria OTU_13 Rhodobacter sp. (AB017799) OTU_12 Rhodobacter blasticus (DQ342322) OTU_10 Rhodobacter blasticus (AM690347) Rhodobacter capsulatus (D16427) OTU_11 OTU_16 Unidentified Clostridiales (EU864476) Unidentified Clostridia (AY953192) Clostridia OTU_2 Unidentified Clostridia (EU234230) OTU_9 Propionicimonas sp. (EF515420) Actinobacteria Propionicimonas paludicola (AB078859) OTU_1 Mollicutes Acholeplasma sp. (EU517562) OTU_8 Unidentified Bacteria (AY945866) OTU_7 E. coli

0.1

Fig. 1 e Phylogenetic tree of the OTUs identiﬁed from the Phototrophic Hydrogen-Producing Bacterial Consortium and their closest relatives, based on partial 16S rRNA gene sequences. The phylogenetic tree was constructed using neighbor-joining algorithm with 100 bootstraps. Escherichia coli was selected as the outgroup species. The scale bar represents 0.1 substitutions per nucleotide position (C, OTUs). international journal of hydrogen energy 37 (2012) 11691e11700 11695

10 as electron donors for sulfate reduction [28]. In this study, the 9 8 presence of acetic acid and H2, which are byproducts of pho- 7 totrophic metabolism, combined with the presence of sulfate, 6 5 may have enabled the growth of these microorganisms. 4 According to Scoma et al. [36] side-process as acetate 3 Number ofNumber clones 2 consumption by sulfate-reducing bacteria can occur when 1 a microbial consortium is used. 0 1 2 3 4 5 6 7 8 9 1011121314151617 The species of interest for this study showed a wide OTU morphological and metabolic diversity and were affiliated to the phototrophic purple non-sulfur bacteria group. These Fig. 2 e Number of clones grouped within each OTU. bacteria belong to Proteobacteria, the largest phylum within the Bacteria. The OTUs containing these microorganisms will be discussed below. OTU_3 was phylogenetically similar to Rhodoplanes cryptolactis (AB087718 e 96%), which is a species that preferentially grows by anaerobic photoorganotrophy metabolism, consuming pyruvate, acetate, malate, succinate and butyrate as carbon sources [37]. OTU_4 was similar to R. palustris (AB017261 e 99%), which is commonly described in the literature as hydrogen producer from various organic acids [38]. OTU_5 and OTU_6 contained clones that were similar to Rhodospirillum photometricum (D30777 e 85%). Another species that belongs to the same genus, Rhodospir- Fig. 3 e Percentage of clones and their phylogenetic illum rubrum, is a purple non-sulfur photosynthetic bacterium affiliations. that can produce hydrogen from various organic acids [39], and also by a water-gas shift reaction [40]. OTU_10 contained 2 clones and OTU_11 contained 3 clones dicarboxylicus gen. nov., sp. nov. a mesophilic Gram-negative similar to R. capsulatus (D16427 e 98%). OTU_12 contained 3 strict anaerobe that was able to grow on maleate, fumarate clones similar to Rhodobacter sp. (AB017799 e 93%). These and malate, producing propionate and acetate as end prod- microorganisms are able to use various organic acids as ucts. The metabolic features of the members of this genus substrates for hydrogen production and thus could have been could explain their presence in the microbial consortium. responsible for generating hydrogen in this study. OTU_15 contained 3 clones similar to Sulfurospirillum sp. Fig. 3 shows the microbial diversity related to the (AY780560 e 92%), and OTU_17 contained 9 clones similar to percentage of clones belonging to different class: Alphapro- Sulfurospirillum cavolei (AB246781 e 100%), which are Gram- teobacteria, Betaproteobacteria, Epsilonproteobacteria and negative, slightly curved rods that can grow fermentatively Clostridia. Thirty-three percent of the clones were classified as on fumarate, pyruvate and malate. The optimum growth of part of the class Alphaproteobacteria, which includes photo- these microorganisms occurs at 30 C and pH 7 [28]. The trophic purple non-sulfur bacteria. Seven percent of the conditions used in this study, including temperature, carbon clones were related to Betaproteobacteria, which also source and pH, were favorable to their growth; however, these contains some species of phototrophic bacteria. Clones microorganisms may not be responsible for the H2 production. similar to Epsilonproteobacteria represented 29% of the total Although they are not responsible for the hydrogen produc- and were similar to the genus Sulfurospirillum. Fourteen tion, they are able to use acetic acid, lactic acid and hydrogen percent of the clones were phylogenetically related to

Table 1 e Summary of the results of hydrogen production experiments from ﬁve different carbon sources. Acetate Butyrate Citrate Lactate Malate

Initial substrate concentration (mmol l 1) 30 17 11 23 14.5 Substrate utilization (%) 43 37 100 49 100 Byproducts-acetic acid (mmol) e 1.4 3.5 1.2 0.8 1 H2 yield (mmol H2 mmol substrate) 0.6 1.4 0.7 0.5 0.9 Substrate efﬁciency conversion (%) 14.5 13.8 8.2 8.2 15.6 1 Cumulative hydrogen production (mmol H2 l culture) 7.8 9.0 7.9 5.6 13.9 aHydrogen production potential 8.8 8.7 8.1 5.9 13.2 1 (mmol H2 l culture) aLag-phase (days) 2.8 2.9 0.5 e 2.2 a 1 1 Maximum rate of hydrogen production (ml H2 l culture h ) 0.4 0.9 0.6 0.3 1.0 aR2 0.991 0.993 0.989 0.984 0.993

a Values obtained by adjusting the data according to the modiﬁed Gompertz equation. 11696 international journal of hydrogen energy 37 (2012) 11691e11700

unidentified Clostridia. Other bacteria were found in smaller proportions: 2% of the clones belong to class Mollicutes, 5% 9 14 were similar to Actinobacteria, and 10% were related to 8 12 culture) 7 -1 unidentified bacteria. 10 ) -1 According to these results we concluded that the set of 6 5 8 experimental conditions (pH, temperature, source of light, 4 carbon source and substrate concentration) were not feasible 6 3 for the enrichment of microbial consortium predominantly 4 Acetate (mmoll 2 composed by phototrophic microorganisms. Solely one third 2 1 of clones were related to this group of bacteria. Another third A 0 0 of clones were related to Sulfurospirillum, which are not l (mmol production Hydrogen 051015202530 desirable microorganisms for the photo-fermentative Time (d) hydrogen production. The high percentage of these microor- 12 20 ganisms (29%) may be related with the low efficiency of the 18

culture) 10 16 process. -1 )

14 -1 8 12 3.2. Hydrogen production from different individual 6 10 carbon sources 8 rate (mmol l

4 y 6 4 But The PHPBC was capable of metabolizing all of the ﬁve different 2 substrates tested in order to produce hydrogen. However, the B 2 0 0 Hydrogen production (mmol l (mmol production Hydrogen percentage of the consumption varied among the substrates. 0 5 10 15 20 25 30 Citrate and malate were totally consumed, while acetate Time (d) (43%), butyrate (37%) and lactate (49%) were not (Table 1). For 12 C 14 the substrates acetate, butyrate and malate, the production of 10 12 culture) -1 hydrogen gas began after two days of incubation, for citrate 10 ) 8 -1 and lactate the production started earlier. The hydrogen 8 production stabilized after approximately 25 days of incuba- 6 6 tion for all the substrates tested. Interestingly, hydrogen 4 production continued even after citrate and malate were 4 l (mmol Citrate completely consumed (Fig. 4). It is possible that the acetic acid 2 2 produced (data not shown) in these reactors may have been 0 0 used as a substrate to maintain hydrogen production. l (mmol production Hydrogen 0 5 10 15 20 25 30 The PHPBC consumed 43% of acetate and produced Time (d) 1 7 25 7.8 mmol H2 l culture. For butyrate, 37% of the substrate was 1 6 consumed and hydrogen production was 9 mmol H2 l culture) 20 -1 )

culture. In this experiment, acetic acid concentrations 5 -1 increased during the incubation period, reaching a maximum 4 15 value of 1.4 mmol of acetic acid after 28 days of incubation. 3 The citrate was totally consumed in 4 days of incubation, and 10 1 2 l (mmol Lactate the cumulative hydrogen production was 7.9 mmol H2 l 5 1 culture. Acetic acid was generated simultaneously with citrate D consumption. At the peak of production (the fourth day), the 0 0 Hydrogen production (mmol l (mmol production Hydrogen acetic acid content reached a concentration of 6.0 mmol, 0 5 10 15 20 25 30 which dropped to 3.5 mmol by the twenty-ﬁfth day of incu- Time (d) 18 16 bation. Probably, the acetic acid generated was most likely E 16 14 also used as a substrate. A total of 49% of lactate was culture) -1 14 consumed after 28 days of incubation, and the acetic acid 12

12 ) concentration reached 1.2 mmol by the end of the experiment. 10 -1 10 The cumulative hydrogen production for this substrate was 8 8 1 6 5.6 mmol H2 l culture. The malate was consumed in its 6 entirety in 9 days of incubation. In this case, acetic acid 4

4 l (mmol Malate concentrations reached a maximum value of 2.5 mmol on the 2 2 fourth day of incubation. The concentration of acetic acid at 0 0 Hydrogen production (mmol l (mmol Hydrogenproduction the end of the experiments was 0.8 mmol. The highest 0 5 10 15 20 25 30 1 Time (d) generation of hydrogen observed was 13.9 mmol H2 l culture in the reactors fed with malate (Table 1). e C L1 Fig. 4 ( ) Cumulative hydrogen production (mmol H2 l By means of a modiﬁed Gompertz equation, the hydrogen culture) and (-) substrate concentration (mM) for (A) Acetate; production potential was determined and the results were (B) Butyrate; (C) Citrate; (D) Lactate; (E) Malate. statistically signiﬁcant, as indicated by the R2 values shown in international journal of hydrogen energy 37 (2012) 11691e11700 11697

Table 1. Thus, this model accurately predicted the cumulative et al. [38] found substrate conversion efficiencies of 14.8%, 0%, phototrophic production of hydrogen. 12.6% and 36%, respectively. The authors tested another Table 2 presents the results obtained in the present work inoculum, which they called the Microbiology strain, with the (substrate conversion efficiency and maximum hydrogen same substrates, and observed substrate conversion efficien- production rate) and also data from other researches to enable cies of 35.3%, 0.3%, 14.4% and 36.7%. In the present study, the an easier comparison among the studies. A wide range of values found for acetate (30 mmol l 1), butyrate (17 mmol l 1), values of substrate conversion efficiency was reported by lactate (23 mmol l 1) and malate (14 mmol l 1) were 14.5%, different authors. It could be due to the various experimental 13.8%, 8.2% and 15.6%, respectively. These values are close to conditions used in each study such as microbial strain, carbon those found by Barbosa et al. [38] for R. palustris, except for the source and substrate concentration [41]. substrates malate and butyrate. Barbosa did not observe any Working with a pure culture of R. palustris and using hydrogen production by R. palustris using butyrate as acetate (22 mmol l 1), butyrate (27 mmol l 1), lactate a substrate, and for malate, the substrate efficiency conver- (50 mmol l 1) and malate (15 mmol l 1) as substrates, Barbosa sion was higher than that in the present study. For the

Table 2 e Comparison of hydrogen production by different microorganisms. Microorganisms Carbon source Initial Maximum rate of Substrate conversion Reference concentration H2 production efﬁciency (%) (mmol l 1) (ml l 1 h 1)

Rhodobacter Acetate 30 20 33 [4] sphaeroides O.U.001 Butyrate 15 5 14 Lactate 20 20 31 Propionate 20 22 31 Malate 15 24 50 Rhodopseudomonas Acetate 22 2.2 14.8 [38] palustris Butyrate 27 ee Lactate 50 9.1 12.6 Malate 15 5.8 36 Microbiology strain Acetate 22 5.3 35.3 Butyrate 27 0.2 0.3 Lactate 50 7.9 14.4 Malate 15 6.0 36.7 ZX-5strain Acetate 35 90 69 [43] Butyrate 50 110 74.6 Succinate 50 94 81.4 Malate 30 92 78.9 Rhodobacter sphaeroides Malate 15 6.9 70.5 [45] O.U.001 Rhodobacter sphaeroides Malate 15 9.2 85.2 O.U.001 hup-mutant Rhodobacter sphaeroides Acetate 30 e 4.6 [46] O.U.001 Malate 15 e 58 Rhodobacter sphaeroides Acetate 30 e 5.5 ZK1 Malate 15 e 71.5 Rhodobacter sphaeroides Acetate 25 26.9 19.5 [41] SCJ Rhodobacter sphaeroides Butyrate 25 31.8 8 SCJ Rhodopseudomonas Acetate 31.4 32.4 73.3 [47] palustris WP3-5 Rhodopseudomonas sp. Butyrate, acetate, 24.5, 18.3, 26.5 56.5 [44] R. palustris W004 propionate, ethanol 2.8, 12.1 10.3 51.6 Rubrivivax sp. 16.4 7.7 Rhodobacter sphaeroides Acetate 60 e 17.5 [48] KD131 Butyrate 30 e 24.7 Rhodobacter sphaeroides Acetate 60 e 36.25 P1 Butyrate 30 e 45.9 PHPBC Acetate 30 0.4a 14.5 Present study Butyrate 17 0.9a 13.8 Citrate 11 0.6a 8.2 Lactate 23 0.3a 8.2 Malate 14 1.0a 15.6

1 1 a Maximum rate of H2 production (mmol H2 l culture d ). 11698 international journal of hydrogen energy 37 (2012) 11691e11700

Microbiology strain, the substrate conversion efficiency was might reduce the substrate conversion efficiency by higher than that observed for R. palustris and also higher than consuming the substrate or even the H2 produced. Thus, more the conversion efficiencies documented in the present study, studies are necessary to evaluate the feasibility of using Pho- except for butyrate. totrophic Hydrogen-Producing Bacterial Consortium instead He et al. [42] reported 52.7% and 68.2% substrate conversion of pure cultures to produce hydrogen gas. efficiency using lactate (30 mmol l 1)byR. capsulatus JP91 and by R. capsulatus IR3 at 30 C, respectively. Tao et al. [43], using RCVB medium at 30 C, isolated a purple non-sulfur bacterium (ZX-5) similar to R. sphaeroides (99.7% similarity) that showed Acknowledgments substrate conversion efficiencies of 69% for acetate (35 mmol l 1), 74.6% for butyrate (50 mmol l 1), 78.9% for The authors wish to thank Saõ Paulo Research Foundation malate (30 mmol l 1) and 81.4% for succinate (50 mmol l 1). (FAPESP) for the financial support of this study and National Uyar et al. [4] showed that R. sphaeroides O.U.001 growing in Council for Scientific and Technological Development (CNPq- Biebl and Pfenning medium and using acetate (30 mmol l 1), Process number 131296/2007-8) for the postgraduate butyrate (15 mmol l 1), lactate (20 mmol l 1) or malate scholarship. (15 mmol l 1) as a carbon source and sodium glutamate (10 mmol l 1) as nitrogen source exhibited substrate conversion efficiencies of 33%, 14%, 31% and 50%, respectively. The references values are higher than those obtained in the present study. Wu et al. [44] studied the hydrogen production by Rhodop- seudomonas sp., R. palustris W004 and Rubrivivax sp. by indi- [1] Das D, Veziroglu TN. Advances in biological hydrogen vidual and mixed carbon source. The authors concluded that production processes. International Journal of Hydrogen Energy 2008;33:6046e57. butyrate was more desirable substrate to produce hydrogen [2] Fascetti E, D’Addario E, Todini O, Robertiello A. than acetate. Furthermore, Rhodopseudomonas sp. produced Photosynthetic hydrogen evolution with volatile organic 1 more hydrogen (2192 ml l culture) than R. palustris W004 acids derived from the fermentation of source selected 1 (1086 ml l culture). Moreover, using the mixture of acetate, municipal solid wastes. International Journal of Hydrogen butyrate, propionate and ethanol, Rhodopseudomonas sp. Energy 1998;23:753e60. showed the higher substrate conversion efficiency (56.5%) in [3] Koku H, Eroglu I, Gunduz U, Yucel M, Turker L. Aspects of the comparison to R. palustris W004 (51.6%) and Rubrivivax sp. metabolism of hydrogen production by Rhodobacter sphaeroides. International Journal of Hydrogen Energy 2002; (7.7%). 27:1315e38. Values of substrate conversion efficiency obtained by [4] Uyar B, Eroglu I, Yucel M, Gunduz U. Photofermentative e [4,43,45 48] are higher than those obtained in the present hydrogen production from volatile fatty acids present in dark study, which is most likely due to the use of pure cultures fermentation effluents. International Journal of Hydrogen instead of a bacterial consortium. According to Valdez- Energy 2009;34:4517e23. Vazquez and Poggi-Varaldo [49], in most anaerobic environ- [5] Zhang T, Liu H, Fang HHP. Microbial analysis of a phototrophic sludge producing hydrogen from acidified ments, H2 is consumed very quickly by different microbial wastewater. Biotechnology Letters 2002;24:1833e7. groups. Moreover, microorganisms present in the microbial [6] Li RY, Fang HHP. Heterotrophic photo fermentative hydrogen consortium could compete with the phototrophic purple non- production. Critical Reviews in Environmental Science and sulfur bacteria by the substrates, causing the decrease in the Technology 2009;39:1081e108. yield of the H2 production process. [7] Sasikala C, Ramana CV, Rao PR. Regulation of simultaneous hydrogen photoproduction during growth by ph and glutamate in Rhodobacter sphaeroides O.U.001. International e 4. Conclusions Journal of Hydrogen Energy 1995;20:123 6. [8] Zhu H, Fang HHP, Zhang T, Beaudette LA. Effect of ferrous ion on photo heterotrophic hydrogen production by Rhodobacter The hydrogen production by the enriched microbial consor- sphaeroides. International Journal of Hydrogen Energy 2007; tium was not feasible considering the low efficiency of the 32:4112e8. processes in terms of hydrogen yield. Although the use of [9] Uyar B, Schumacher M, Gebicki J, Modigell M. microbial consortium is a crucial point for scale up the bio- Photoproduction of hydrogen by Rhodobacter capsulatus from logical hydrogen production, it can create a number of prob- thermophilic fermentation effluent. Bioprocess and Biosystems Engineering 2009;32:603e6. lems due to the concomitant presence of H producers and 2 [10] Adessi A, Torzillo G, Baccetti E, De Philippis R. Sustained consumers. Physiological parameters such as pH, tempera- outdoor H2 production with Rhodopseudomonas palustris ture, and substrate, as well as, process parameters as cultures in a 50 L tubular photobioreactor. International hydraulic retention time and organic loading rate, could be Journal of Hydrogen Energy 2012;37:8840e9. adjusted in reactors operating in a continuous mode to favor [11] Oh YK, Seol EH, Kim MS, Park S. Photoproduction of some microorganisms in spite of others. Based on molecular hydrogen from acetate by a chemoheterotrophic bacterium Rhodopseudomonas palustris biology analysis, the consortium was composed of photo- P4. International Journal of Hydrogen Energy 2004;29:1115e21. trophic purple non-sulfur bacteria, which may be responsible [12] Chen CY, Lu WB, Liu CH, Chang JS. Improved phototrophic H2 for the photobiological production of H2. However, other production with Rhodopseudomonas palustris WP3-5 using microorganisms, including Sulfurospirillum, were also found, acetate and butyrate as dual carbon substrates. Bioresource e which are not related to H2 production, and furthermore Technology 2008;99:3609 16. international journal of hydrogen energy 37 (2012) 11691e11700 11699

[13] Eroglu E, Gunduz U, Yucel M, Eroglu I. Photosynthetic [31] Maintinguer SI, Fernandes BS, Duarte ICS, Saavedra NK, bacterial growth and productivity under continuous Adorno MAT, Varesche MBA. Fermentative hydrogen illumination or diurnal cycles with olive mill wastewater as production by microbial consortium. International Journal of feedstock. International Journal of Hydrogen Energy 2010;35: Hydrogen Energy 2008;33:4309e17. 5293e300. [32] Maintinguer SI, Fernandes BS, Duarte ICS, Saavedra NK, [14] Seifert K, Waligorska M, Laniecki M. Hydrogen generation in Adorno MAT, Varesche MBA. Fermentative hydrogen photobiological process from dairy wastewater. production with xylose by Clostridium and Klebsiella species in International Journal of Hydrogen Energy 2010;35:9624e9. anaerobic batch reactors. International Journal of Hydrogen [15] Kapdan IK, Kargi F, Oztekin R, Argun H. Bio-hydrogen Energy 2011;36:13508e17. production from acid hydrolyzed wheat starch by photo- [33] Akasaka H, Ueki A, Hanada S, Kamagata Y, Ueki K. fermentation using different Rhodobacter sp. International Propionicimonas paludicola gen. nov., sp. nov., a novel Journal of Hydrogen Energy 2009;34:2201e7. facultatively anaerobic, gram-positive, propionate- [16] Avcioglu SG, Ozgur E, Eroglu I, Yucel M, Gunduz U. producing bacterium isolated from plant residue in irrigated Biohydrogen production in an outdoor panel photobioreactor rice-field soil. International Journal of Systematic and on dark fermentation effluent of molasses. International Evolutionary Microbiology 2003;53:1991e8. Journal of Hydrogen Energy 2011;36:11360e8. [34] Garrity GM. The proteobacteria: part C the alpha-, beta-, [17] Afsar N, Ozgur E, Gurgan M, Akkose S, Yucel M, delta-, and Epsilonproteobacteria. Genus IV. Gunduz U, et al. Hydrogen productivity of photosynthetic Magnetospirillum. In: Bergey’s manual of systematic bacteria on dark fermenter effluent of potato steam peels bacteriology. 2nd ed. New York: Springer; 2005. hydrolysate. International Journal of Hydrogen Energy [35] Tanaka K, Nakamura K, Mikami E. Fermentation of maleate 2011;36:432e8. by a gram-negative strictly anaerobic non-spore-former, [18] Weaver PF, Wall JD, Gest H. Characterization of Propionivibrio dicarboxylicus gen. nov., sp. nov. Archives of Rhodopseudomonas capsulata. Archives of Microbiology 1975; Microbiology 1990;154:323e8. 105:207e16. [36] Scoma A, Bertin L, Zanaroli G, Fraraccio S, Fava F. A [19] Fang HHP, Liu H, Zhang T. Phototrophic hydrogen production physicochemical-biotechnological approach for an from acetate and butyrate in wastewater. International integrated valorization of olive mill wastewater. Bioresource Journal of Hydrogen Energy 2005;30:785e93. Technology 2011;102:10273e9. [20] APHA AWWA WEF. Standard methods for the examination [37] Okamura K, Hisada T, Hiraishi A. Characterization of of water and wastewater. 20th ed. Washington, DC, USA: thermotolerant purple non-sulfur bacteria isolated from hot- APHA; 1998. spring Chloroflexus mats and the reclassification of [21] Griffiths RI, Whiteley AS, O’Donnell AG, Bailey MJ. Rapid “Rhodopseudomonas cryptolactis” Stadtwald-Demchick et al. method for coextraction of DNA and RNA from natural 1990 as Rhodoplanes cryptolactis nom. rev., comb. nov. Journal environments for analysis of ribosomal DNA- and rRNA- of General and Applied Microbiology 2007;53:357e61. based microbial community composition. Applied and [38] Barbosa MJ, Rocha JMS, Tramper J, Wijffels RH. Acetate as Environmental Microbiology 2000;66:5488e91. a carbon source for hydrogen production by photosynthetic [22] Lane DJ. 16S/23S rRNA sequencing. In: Stackebrandt E, bacteria. Journal of Biotechnology 2001;85:25e33. Goodfellow M, editors. Nucleic acid techniques in bacterial [39] Zhu R, Li J. Hydrogen metabolic pathways of Rhodospirillum systematics. Chichester, United Kingdom: John Wiley & rubrum under artificial illumination. Chinese Science Bulletin Sons; 1991. p. 115e75. 2010;55:32e7. [23] Chun J. Computer assisted classification and identification of [40] Klasson KT, Cowger JP, Ko CW, Vega JL, Clausen EC, Gaddy JL. Actinomycetes. Ph.D. thesis. Newcastle upon Tyne, UK: Methane production from synthesis gas using a mixed University of Newcastle upon Tyne; 1995. culture of R. rubrum, M. barkeri, and M. formicicum. Applied [24] Oliveira LL, Costa RB, Okada DY, Vich DV, Duarte ICS, Biochemistry and Biotechnology 1990;24-25:317e28. Silva EL, et al. Anaerobic degradation of linear alkylbenzene [41] Lee JZ, Klaus DM, Maness PC, Spear JR. The effect of butyrate sulfonate (LAS) in fluidized bed reactor by microbial concentration on hydrogen production via consortia in different support materials. Bioresource photofermentation for use in a Martian habitat resource Technology 2010;101:5112e22. recovery process. International Journal of Hydrogen Energy [25] Saitou N, Nei M. The neighbor-joining method e a new 2007;32:3301e7. method for reconstructing phylogenetic trees. Molecular [42] He DL, Bultel Y, Magnin JP, Willison JC. Kinetic analysis of Biology and Evolution 1987;4:406e25. photosynthetic growth and photohydrogen production of [26] Lay JY, Li YY, Noike T. Developments of bacterial population two strains of Rhodobacter capsulatus. Enzyme and Microbial and methanogenic activity in a laboratory-scale landfill Technology 2006;38:253e9. bioreactor. Water Research 1998;32:3673e9. [43] Tao YZ, He YL, Wu YQ, Liu FH, Li XF, Zong WM, et al. [27] Imhoff JF. The phototrophic alpha-proteobacteria. In: Characteristics of a new photosynthetic bacterial strain for Dworkin M, et al., editors. The prokaryotes, vol. 5, part 1. hydrogen production and its application in wastewater New York: Springer; 2006. p. 41e64. treatment. International Journal of Hydrogen Energy 2008;33: [28] Kodama Y, Ha LT, Watanabe K. Sulfurospirillum cavolei sp nov, 963e73. a facultatively anaerobic sulfur-reducing bacterium isolated [44] Wu X, Wang X, Yang H, Guo L. A comparison of from an underground crude oil storage cavity. International hydrogen production among three photosynthetic Journal of Systematic and Evolutionary Microbiology 2007;57: bacterial strains. International Journal of Hydrogen 827e31. Energy 2010;35:7194e9. [29] Hill AC. Acholeplasma cavigenitalium sp. nov., isolated from [45] Kars G, Gunduz U, Rakhely G, Yucel M, Eroglu I, the vagina of guinea-pigs. International Journal of Kovacs KL. Improved hydrogen production by uptake Systematic Bacteriology 1992;42:589e92. hydrogenase deficient mutant strain of Rhodobacter [30] Fang HHP, Zhu HG, Zhang T. Phototrophic hydrogen sphaeroides O.U.001. International Journal of Hydrogen production from glucose by pure and co-cultures of Energy 2008;33:3056e60. Clostridium butyricum and Rhodobacter sphaeroides. [46] Kars G, Gunduz U, Yucel M, Rakhely G, Kovacs KL, Eroglu I. International Journal of Hydrogen Energy 2006;31:2223e30. Evaluation of hydrogen production by Rhodobacter sphaeroides 11700 international journal of hydrogen energy 37 (2012) 11691e11700

O.U.001 and its hupSL deﬁcient mutant using acetate and [48] Kim M-S, Kim D-H, Son H-N, Ten LN, Lee JK. Enhancing malate as carbon sources. International Journal of Hydrogen photo-fermentative hydrogen production by Rhodobacter Energy 2009;34:2184e90. sphaeroides KD131 and its PHB synthase deleted-mutant from [47] Chen C-Y, Saratale GD, Lee C-M, Chen P-C, Chang J-S. acetate and butyrate. International Journal of Hydrogen Phototrophic hydrogen production in photobioreactors Energy 2011;36:13964e71. coupled with solar-energy-excited optical ﬁbers. [49] Valdez-Vazquez I, Poggi-Varaldo HM. Hydrogen production International Journal of Hydrogen Energy 2008;33: by fermentative consortia. Renewable & Sustainable Energy 6886e95. Reviews 2009;13:1000e13.