Enrichment of syngas‐converting communities from a multi‐orifice baffled bioreactor

Summary The substitution of natural gas by renewable biomethane is an interesting option to reduce global carbon footprint. Syngas fermentation has potential in this context, as a diverse range of low‐biodegradable materials that can be used. In this study, anaerobic sludge acclimatized to syngas in a multi‐orifice baffled bioreactor (MOBB) was used to start enrichments with CO. The main goals were to identify the key players in CO conversion and evaluate potential interspecies metabolic interactions conferring robustness to the process. Anaerobic sludge incubated with 0.7 × 105 Pa CO produced methane and acetate. When the antibiotics vancomycin and/or erythromycin were added, no methane was produced, indicating that direct methanogenesis from CO did not occur. Acetobacterium and Sporomusa were the predominant bacterial species in CO‐converting enrichments, together with methanogens from the genera Methanobacterium and Methanospirillum. Subsequently, a highly enriched culture mainly composed of a Sporomusa sp. was obtained that could convert up to 1.7 × 105 Pa CO to hydrogen and acetate. These results attest the role of Sporomusa species in the enrichment as primary CO utilizers and show their importance for methane production as conveyers of hydrogen to methanogens present in the culture.


Introduction
Biomass can be biologically converted to fuels and chemicals, but poorly biodegradable biomass (e.g. straw, wood) requires costly chemical or enzymatic hydrolysis prior to microbial fermentation. Gasification of biomass, and subsequent syngas fermentation, is an alternative that can maximize carbon recovery compared to conventional fermentations (Bredwell et al., 1999;Munasinghe and Khanal, 2010;Guiot et al., 2011). Syngas is mainly composed of CO, H 2 and CO 2, and these compounds can be used by anaerobic microorganisms to produce value-added chemicals. In the absence of external electron acceptors, CO can be converted by hydrogenogenic, acetogenic and methanogenic microorganisms (Diender et al., 2015). The hydrogenogenic conversion of CO results in the formation of H 2 and CO 2 , and is typically performed by thermophilic carboxydotrophs, such as Carboxydothermus hydrogenoformans, Moorella stamsii and Desulfotomaculum nigrificans (Svetlitchnyi et al., 2001;Parshina et al., 2005;Henstra et al., 2007;Alves et al., 2013a;Visser et al., 2014;Diender et al., 2015). Carboxydotrophic mesophiles, such as Clostridium carboxidivorans, C. ljungdahlii, C. autoethanogenum, Acetobacterium woodii, Sporomusa ovata and Butyribacterium methylotrophicum, have the capability of converting CO into short-chain fatty acids (mainly acetate) and alcohols (ethanol, butanol, 2,3-butanediol) (Henstra et al., 2007;Balk et al., 2010;K€ opke et al., 2010Munasinghe and Khanal, 2011;Jang et al., 2012;Liu et al., 2014a,b;Diender et al., 2015). In general, CO is a poor substrate for methanogens. Methane production from CO has been reported only for some species of Methanosarcina and Methanothermobacter (Daniels et al., 1977;Rother and Metcalf, 2004;Henstra et al., 2007;Diender et al., 2016). For this reason, the utilization of mixed cultures (sludges) for the production of methane from syngas may be advantageous as normally these systems are more robust and less susceptible to inhibition (Guiot et al., 2011). Although methane production directly from CO is possible, hydrogenotrophic and acetoclastic methanogens can coexist with carboxydotrophic, hydrogenogenic and acetogenic bacteria and use the bacterial products (H 2 and acetate) to ultimately produce methane. An additional advantage of using anaerobic sludge for syngas conversion is the possibility of implementing low-cost, open fermentation This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. bs_bs_banner systems. Methane is a direct substitute of natural gas, and infrastructure to distribute methane to industry and households is existing.
Bioreactor technology for open fermentation of syngas is also developing, with special attention to the requisites of high gas-liquid mass transfer rates and cell retention times. Several reactor types have been studied for syngas open fermentations, including gas-lift reactors (Haddad et al., 2014), reverse membrane bioreactors (Youngsukkasem et al., 2015;Westman et al., 2016) and multi-orifice baffled bioreactor (MOBB) (Pereira, 2014). The MOBB described by Pereira (2014) showed methane production rates of about 1.5-2 times higher than reported for the other systems. In this study, we analysed the microbial communities in the MOBB sludge and performed enrichment and microbial diversity studies to get more insight into key players and metabolic networks involved in CO biomethanation. The main aim was to verify whether methane production from CO occurs predominantly via direct methanogenesis or indirectly via bacterial-archaeal associations. Identifying key microbial interactions could explain and justify the higher robustness of mixed cultures for syngas conversion to methane.

Results and discussion
Multi-orifice baffled (bio)reactors (MOBB) are recognized for their excellent performance in gas-liquid mass transfer and therefore suited for the conversion of gaseous substrates (Ni et al., 2003). In a previous work, Pereira (2014) described the application of a MOBB to continuous syngas conversion to methane using open anaerobic mixed cultures (sludge). Although methane was efficiently produced from CO, the microorganisms involved in this conversion and possible metabolic interactions were not disclosed. In this study, we used a combination of enrichment studies and microbial diversity analyses to identify key microbial players and microbial interactions occurring in the MOBB sludge.

CO to methane conversion by MOBB enrichments is dependent on bacterial-archaeal interactions
Incubation of MOBB sludge with 40% CO as sole carbon and energy source resulted in the production of methane and acetate. In Fig. 1, substrate consumption and product formation are shown for a stable enrichment culture obtained after 12 successive transfers of the MOBB sludge on CO (CO(12)): 68.9 mmol L À1 medium CO resulted in the formation of 10.4 mmol L À1 medium methane and 7.9 mmol L À1 medium acetate. When the MOBB sludge was incubated in the presence of the antibiotics vancomycin and erythromycin (CO-v 50 e 100 and O-CO-v 100 e 100 ), no CO conversion or methane production was observed over a long incubation period (over 3 months) (data not shown). Vancomycin and erythromycin are bacterial inhibitors and the fact that no CO conversion was observed in their presence indicates that CO was not directly metabolized by carboxydotrophic methanogens. Carbon monoxide is inhibitory for most methanogens, and only a few species can actually grow on this substrate: Methanothermobacter thermoautotrophicus (Daniels et al., 1977), Methanothermobacter marburguensis (Diender et al., 2016), Methanosarcina barkeri (O'Brien et al., 1984;Bott et al., 1986) and Methanosarcina acetivorans (Rother and Metcalf, 2004). Both thermophilic Methanothermobacter species grow significantly slower on CO than on H 2 /CO 2 (Daniels et al., 1977;Diender et al., 2016). Methanosarcina acetivorans can withstand higher CO partial pressures than the hydrogenotrophic methanogens, but its methanogenic metabolism shifts towards formation of acetate and formate at increased CO pressures (Rother and Metcalf, 2004).
Analysis of the archaeal communities in the inoculum and initial CO enrichment cultures by cloning and sequencing (Sanger) revealed the presence of hydrogenotrophic methanogens closely related to Methanobacterium and Methanospirillum species (Fig. 2). However, microorganisms related to known carboxydotrophic methanogens were not detected, which again supports the hypothesis that methane was not directly produced from CO in these cultures. Methanobacterium and Methanospirillum species are capable of using H 2 /CO 2 to produce methane (Whitman et al., 2006). Although H 2 could not be detected in CO enrichments (Fig. 1), microbial H 2 production from CO is possible (water-gas shift reaction, CO + H 2 O ? H 2 + CO 2 ) (Diender et al., 2015). The two predominant bacteria identified in the enrichments were related to Acetobacterium and Sporomusa species (Fig. 2), which normally have a homoacetogenic metabolism when growing on CO (Genthner and Bryant, 1987;Balk et al., 2010;Diender et al., 2015). However, interspecies H 2 transfer during growth of Acetobacterium woodii with different methanogens on sugars has been previously reported (Winter and Wolfe, 1980), indicating that a similar mechanism could take place during CO fermentation. Acetoclastic methanogens were not detected in the enrichment cultures (Fig. 2), and acetate accumulation was observed (Fig. 1). Methanosarcina and Methanosaeta species are commonly present in anaerobic sludges (Karakashev et al., 2005;Pan et al., 2016;Sancho Navarro et al., 2016), and the reason for their absence in the MOBB sludge is not clear. It might be that these microorganisms suffer from the shear stress caused by the oscillations in the MOBB. In addition, effects of CO toxicity towards aceticlastic methanogens cannot be excluded.
A novel Sporomusa sp. is present in MOBB sludge that can convert CO to H 2 From the DGGE profiles of cultures CO(0) and CO(1), it is evident that bacterial communities were dominated by Acetobacterium and Sporomusa species (Fig. 2). Carbon monoxide metabolism is well studied in Acetobacterium woodii. This organism was initially described to grow homoacetogenically on CO as a sole energy source (Genthner and Bryant, 1987), but it has been recently found that it can use CO only in cofermentation with H 2 or formate (Bertsch and M€ uller, 2015). Acetobacterium population in the CO enrichments is most closely related to Acetobacterium wieringae (99% 16S rRNA gene identity). Recently, Sancho Navarro et al. (2016) reported the presence of bacterial species closely related to A. wieringae after incubation of anaerobic sludge at high CO concentrations. While only A. woodii is reported as carboxydotrophic organism, A. wieringae genome contains the gene clusters for both carbonyl and methyl branches of the Wood-Ljungdahl pathway, identical to what is found in A. woodii (Genthner and Bryant, 1987;Poehlein et al., 2016). Acetobacterium population became more predominant during the enrichment process (Fig. S1), and based on MiSeq results, it represented about 82% of the community in enrichment culture CO(12) ( Table 1A).
The presence of closely related organisms to Sporomusa species in the initial enrichments was rather striking. Carbon monoxide metabolism in Sporomusa species is poorly studied, although CO conversion to acetate has been reported for S. termitida, S. ovata and Sporomusa strain An4 (Breznak et al., 1988;Balk et al.,  . Sporomusa strains are known to form spores (M€ oller et al., 1984), and it was clear from microscopic examination that spores were present in early CO enrichment cultures (data not shown). With the aim of enriching/isolating the Sporomusa species in the CO enrichments, we proceeded with the pasteurization of culture CO(4), and later on, a second pasteurization was performed with culture CO-P(4) (see experimental set-up in Fig. 4). Culture CO-P(8) was highly enriched in Sporomusa (97% of the total microorganisms) ( Table 1B) and was able to grow under a 100% CO headspace. Culture CO-P(8)-produced H 2 from CO (Fig. 3), which considering the composition of the initial CO enrichment cultures, could benefit growth of both Acetobacterium species and hydrogenotrophic methanogens. H 2 production by Sporomusa strains from CO was not shown before, but these bacteria have an important role in CO conversion in the mixed culture. A remarkable observation is that the Sporomusa species in CO-P(12) can grow with 1.7 9 10 5 Pa CO. Growth inhibition of S. termitida was observed for CO partial pressures higher than 0.4 9 10 5 Pa (Breznak et al., 1988). Further research is needed to compare the enriched Sporomusa strain with other Sporomusa-type strains.

Source of inoculum and medium composition
Granular sludge obtained from a 10-L MOBB fed with syngas mixture (60% CO, 30% H 2 and 10% CO 2 (v/v)) (Pereira, 2014) was used as inoculum for batch incubations with CO. Cultures were prepared in 120-ml serum bottles containing 30 ml of bicarbonate-buffered mineral salt medium and sealed with butyl rubber stoppers and aluminium crimp caps. Medium was prepared as described by Stams et al. (1993). Carbon monoxide was used as the sole energy and carbon source and added to the bottles' headspace to the desired final partial pressure using a syringe. All the assays were carried out at initial total pressure of 1.7 9 10 5 Pa, for which a mixture of N 2 /CO 2 (80:20 (v/v)) was added to the headspace when necessary. For assays with CO partial pressures higher than 1.3 9 10 5 Pa, a phosphate-buffered mineral salt medium was used and prepared as described by Alves et al. (2013b). Medium was reduced prior to inoculation with 0.8 mM (final concentration) sodium sulphide (Na 2 SÁxH 2 O, x = 7-9). Cultures were monitored by measuring CO depletion and methane, hydrogen and acetate production.

Batch assays
An overview of the experimental set-up implemented in this study is illustrated in Fig. 4. To test the direct methanogenesis from CO, methanogenic sludge was incubated with the antibiotics vancomycin (v) and erythromycin (e) (bacterial growth inhibitors): CO-v 50 e 100 (50 lM vancomycin + 100 lM erythromycin) and COv 100 e 100 (100 lM of each antibiotic). Incubations without  antibiotics were also performed. Bottles' headspace contained 40% CO (pCO = 6.8 9 10 4 Pa) as substrate, and cultures were incubated at 37°C. None of the incubations with antibiotics could use CO or produce methane. Incubations without antibiotics were successively transferred (10% (v/v)) to new bicarbonate-buffered medium and fed with CO for 2 years with 40% CO, resulting in an enrichment series designated as CO(x) (where x stands for number of successive transfers). Since an early enrichment stage, the presence of sporulating cells in the CO cultures was evident (microscopic observation). An aliquot of culture CO(4) was pasteurized and used as inoculum to start the enrichment series CO-P(x) (where P stands for pasteurization; x stands for number of successive transfers). Pasteurization procedure consisted in heating up the culture to 80°C for 20 min. Culture CO-P(1) was transferred four times to fresh medium, and at this time, a second pasteurization was performed (culture CO-P(4)). Pasteurization temperature was now increased to 95°C (for 20 min). In transfers immediately after each pasteurization, yeast extract and ethanol were used to stimulate growth, being removed in posterior transfers. CO was the sole carbon and energy source added (40%; pCO = 6.8 9 10 4 Pa) in the next transfers (CO-P(5) and CO-P(6)). Then, the CO partial pressure in the headspace was raised gradually from 40% (pCO = 6.8 9 10 4 Pa) to 100% CO (pCO = 1.7 9 10 5 Pa) (CO-P(7) to CO-P(12)).

Analytical methods
Gas samples were analysed by gas chromatography with a Bruker Scion 456-GC (Billerica, MA, USA) with a thermal conductivity detector and equipped with two columns: a BR-QPLOT column (30 m length, 0.53 mm internal diameter; film thickness, 20 lm) and a Molsieve packed column (13 9 80/100, 2 m length, 2.1 mm internal diameter). The Molsieve column was used to measure CO, H 2 and CH 4, and argon was used as carrier gas at a flow rate of 30 ml min À1 ; temperatures in the injector, column and detector were 100, 35 and 130°C respectively. Volatile fatty acids (VFA), such as acetate, were determined by high-performance liquid chromatography using an HPLC (Jasco, Tokyo, Japan) with a Phenomenex Rezex ROAorganic acid H+ (8%) column (300 9 7.8 mm). The mobile phase used was sulfuric acid (0.005 N) at a flow rate of 0.6 ml min À1 . Column temperature was set at 60°C. Detection of VFA was made sequentially with an UV detector at 210 nm.

DNA extraction and amplification
Five millilitres of well-homogenized MOBB sludge, CO(x) and CO-P(x) cultures was stored at À20°C. DNA extraction from these samples was performed using the Fas-tDNA SPIN kit for soil (MP Biomedicals, Solon, OH) according to the manufacturer's protocol. The microbial 16S rRNA genes were amplified by PCR using a Taq DNA polymerase kit (Invitrogen, Carlsbad, CA). PCR programs and reactions mixtures used were as described elsewhere (Sousa et al., 2007), and all primers set were synthesized by Invitrogen. For 16S rRNA gene amplification for denaturing gradient gel electrophoresis (DGGE), primers set 1401r/968-GCf was used for bacteria and 515-GCr/A109(T)f for archaea (Sousa et al., 2007). The yield and size of PCR products were assessed by electrophoresis in 1% agarose gel (wt/vol), using a 1 kb extended DNA ladder (ThermoScientific, Waltham, MA, USA) and a green safe staining.

DGGE analysis
DGGE analysis was performed using a DCode system (Bio-Rad, Hercules, CA). For the purpose, gels of 8% (wt/ vol) polyacrylamide (37.5:1 acrylamide/bis-acrylamide) were used with a denaturing gradient of 30% to 50% for archaea and 30% to 60% for bacteria, with 100% of denaturant corresponding to 7 M urea and 40% (v/v) formamide. Electrophoresis ran in a 0.5 TAE buffer at 60°C for 16 h at 85 V. Posteriorly gels were stained with silver nitrate (Sanguinetti et al., 1994) and scanned in an Epson Perfection V750 PRO (Epson, Long Beach, CA, USA).

16S rRNA gene sequencing
Cloning and Sanger sequencing were performed using the methodologies previously described by Sousa et al. (2007). Similarity searches were performed using the NCBI BLAST search program within the GenBank database (http://www.ncbi.nlm.nih.gov/blast/) (Altschul et al., 1990). Illumina Miseq platform sequencing was performed at the Research and Testing Laboratory -RTLGenomics (Lubbock, TX). The MiSeq method used was the Illumina two-step using universal primers for bacteria and archaea, 515f and 806r developed by Caporaso et al. (2011). After sequencing, the data were processed using the data analysis pipeline from RTL, which consists in two major steps, the denoizing and chimera detection step and the microbial diversity analysis step, as described in the company procedures.

Nucleotide sequence accession numbers
The 16S rRNA gene sequences obtained in this study have been deposited in the European Nucleotide Archive (ENA) under the accession numbers LT671598-LT671603 (Sanger sequencing) and ERS1422865-ERS1422866 (Illumina MiSeq platform); these sequences can be viewed by following URL http://www.ebi.ac.uk/ena/data/view/PRJ EB16760).