Analysis of the Microbial Community in an Acidic Hollow-Fiber Membrane Biofilm Reactor (Hf-MBfR) Used for the Biological Conversion of Carbon Dioxide to Methane

Hydrogenotrophic methanogens can use gaseous substrates, such as H2 and CO2, in CH4 production. H2 gas is used to reduce CO2. We have successfully operated a hollow-fiber membrane biofilm reactor (Hf-MBfR) for stable and continuous CH4 production from CO2 and H2. CO2 and H2 were diffused into the culture medium through the membrane without bubble formation in the Hf-MBfR, which was operated at pH 4.5–5.5 over 70 days. Focusing on the presence of hydrogenotrophic methanogens, we analyzed the structure of the microbial community in the reactor. Denaturing gradient gel electrophoresis (DGGE) was conducted with bacterial and archaeal 16S rDNA primers. Real-time qPCR was used to track changes in the community composition of methanogens over the course of operation. Finally, the microbial community and its diversity at the time of maximum CH4 production were analyzed by pyrosequencing methods. Genus Methanobacterium, related to hydrogenotrophic methanogens, dominated the microbial community, but acetate consumption by bacteria, such as unclassified Clostridium sp., restricted the development of acetoclastic methanogens in the acidic CH4 production process. The results show that acidic operation of a CH4 production reactor without any pH adjustment inhibited acetogenic growth and enriched the hydrogenotrophic methanogens, decreasing the growth of acetoclastic methanogens.


Introduction
Most methanogens convert carbon dioxide (CO 2 ) to methane (CH 4 ), the major flammable component of natural gas. CH 4 can be used to make a renewable, carbon-neutral gas substitute [1][2][3]. Hydrogenotrophic methanogens can upgrade CO 2 to CH 4 using molecular hydrogen maintained in anaerobic condition by purging mixed gases (H 2 :CO 2 = 4:1) through the hollow-fiber membrane and by keeping the inner reactor temperature at 35-38°C with a heating circulator. The supplied mixed gases were used as substrates for the biological conversion of CO 2 to CH 4 . The pressure of the mixed gases was 1.4-2.1 kPa. Sieved anaerobic digested sludge (initial inoculum, S1 Fig) collected from a wastewater treatment plant (Jungrang wastewater treatment plant, Seoul) was inoculated into each reactor at a level of 20% of the working volume. The mineral medium composition of the reactor is shown in Table 1. The inner reactor was mixed with the up-flow using a recycling pump. ORP (oxidation-reduction potential), temperature, and pH were continuously observed with an ORP probe, a thermometer, and a pH meter, respectively. A wet gas-meter (Model W-NK-0.5, Shinagawa, Japan) was used to measure the volume of gases produced. Gases taken from the sampling port installed at the gas effluent line were analyzed by GC-TCD, and volatile organic acid analysis was performed by GC-FID.

Total genomic DNA extraction
Samples (3-5 mL) were taken at the recycling line for microbial community analysis. After centrifugation, the supernatant was collected for volatile organic acids analysis, and the pellet was used for the extraction of total genomic DNA with a Power Soil DNA Kit (MO BIO, Carlsbad, USA). The quality of the extracted DNA was examined with standard agarose gel electrophoresis and stored at -20°C. The extracted DNA was used as the template for the 16S rDNA PCR.
Bacterial and archaeal 16S rDNA PCR-DGGE PCR for microbial community analysis was conducted with nested PCR. The 1 st PCR amplification of the bacterial 16S rRNA genes was performed with primers 27f and 1492r, and with primers 46f and 1100r for archaeal 16S rRNA genes. The PCR amplification was performed in 50 μl reaction mixtures containing 5μl Ex 10× PCR buffer (Takara, Japan), 8 μl 2.5 mM dNTPs, 2 μl each primer (20 pmol/μl), 2.5U of Ex Taq polymerase (Takara, Japan), 2 μl template DNA and 30.5 μl distilled water. The thermo cycling program was: 5 min initial denaturation at 95°C (94°C for 2 min for archaeal DNA), followed by 30 cycles of 1 min at 94°C, 1 min at 58°C (57°C for 1 min), and 1 min at 72°C, followed by 10 min of final extension at 72°C. The 2 nd PCR amplification of the 16S rRNA genes was performed with primers with the primer pairs 341f-GC/534r and 340f-GC/519r for bacterial and archaeal DNA, respectively. The PCR For sequencing, the selected bands were excised from the DGGE gels using a sterile scalpel and placed in a sterile Eppendorf tube containing 40 μl of sterile water, and the DNA was eluted using five cycles of freeze-thawing (-70°C/37°C). Two microliters of the solution were used as template DNA in the PCR using the 2 nd bacterial and archaeal PCR protocol with non-GC clamp primers. The amplified products were purified from agarose gel slices with a QIAquick Gel Extraction Kit (QIAGEN, Valencia, CA). The purified PCR products were cloned with the yT&A vector cloning kit (Yeastern Biotech, Taipei, Taiwan) following the manufacturer's instructions. Two or three white colonies were randomly picked from each cloned sample for plasmid isolation (Nucleogen, Daejeon, South Korea). Sequencing was performed with an ABI Prism BigDye Terminator Cycle Sequencing Ready Kit (Applied Biosystems, Foster City, USA) according to the manufacturer's instructions with the sequencing primer M13f for the vector. Approximately 200 unambiguous nucleotide positions were used for comparison with the data in GenBank using the Basic Local Alignment Search Tool (BLAST). Sequences from the nearest relatives were identified from BLAST.

Real-time qPCR for bacteria and archaea
The DNAs extracted from the samples of the reactors were used to construct standard curves for bacteria, archaea, Methanobacteriales and Methanomicrobiales. First, the PCR results using each 16S rRNA primer ( Table 2) were used to determine the copy number after transformation for each taxonomic group. The PCR amplification was performed in 25 μl reaction mixtures containing 2.5μl Ex 10× PCR buffer (Takara, Japan), 4 μl 2.5 mM dNTPs, 2.5 μl of each primer (9 pmol/μl), 1.25 U Ex Taq polymerase (Takara, Japan), 1 μl of template DNA and 12.25 μl of distilled water. The PCR cycles consisted of an initial denaturation at 95°C for 10 min, followed by 30 cycles of 95°C for 30 sec, with annealing temperatures of 58°C, 60°C, 60°C, and 63°C for bacteria, archaea, Methanobacterialses, and Methanomicrobiales, respectively, for 1 min, and 72°C for 30 sec and a final extension at 72°C for 10 min. The 16S rRNA genes with proper sizes were extracted and purified with a QIAquick Gel Extraction Kit (QIAGEN, Valencia, USA). Each purified DNA was cloned with the yT&A vector cloning kit (Yeastern Biotech. Taipei, Taiwan) following the manufacturer's instructions. Transformed colonies were incubated in LB broth (1% tryptone, 0.5% NaCl, and 0.5% yeast extract) at 37°C for 16 h. The culture solution was centrifuged, and plasmids were obtained from the pellet. PCR was performed to obtain a high concentration of DNA as described above, using these plasmids as a template. The concentrations of the PCR products were measured with a BIORAD VersaFluor™ fluorometer. By assuming an average molecular weight of 660 Da for a base pair in double-stranded DNA [18], the following equation (Eq 1) was used to calculate the number of 16S rRNA gene copies that were present in the obtained DNA [19].
16S rDNA ðcopy=mLÞ ¼ 16S rDNA concentration ðg=mlÞ Â 6 Â 10 23 16S rDNA amplicon size ðbpÞ Â 660ðg 16S rDNA=mol=bpÞ ð1Þ Initial 16S rRNA gene copy numbers ranged on the order of 10 11 for each target, and they were serially diluted. The diluted samples and unknown samples were amplified with a realtime qPCR system (Applied Biosystems 7300 Real Time PCR System) with each primer and probe set (Table 3). Real-time qPCR was performed in 25 μlreaction mixtures containing 12.5μl 2×PCR master mix (Promega, Fitchburg, USA), 2.5 μlof each primer (9 pmol/μl) and probe (2.5 pmol/μl), 1 μl template DNA and 4 μl distilled water. PCR cycles consisted of 50°C  Table 3. Primer and probe sets used in this study for qPCR assay. for 2 min and an initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec, and then a combined annealing/extension step at 60°C for 1 min (58°C for bacteria). Logarithmic values of the different 16S rRNA gene amounts were plotted against the threshold cycle (C T ) numbers from each qPCR assay. The linear ranges of the standard curves were selected based on the r 2 of the slope being greater than 0.995 except for Methanomicrobiales (0.930).

Pyrosequencing
Upon attainment of maximum CH 4 production, total DNA was extracted with the Power Soil TM DNA isolation kit (MO BIO, Carlsbad, USA), according to the manufacturer's instructions. The 16S rRNA genes were amplified (Roche 454 GS FLX Titanium) using bar-coded universal primers for each sample. The primer sequences were as follows: bacterial universal (27F: AGA GTT TGA TCM TGG CTC AG, 518r: WTT ACC GCG GCT GCT GG) and archaeal universal (arc112F: GCT CAG TAA CAC GTG G, arc516r: GGT DTT ACC GCG GCK GCT G) for bacterial and archaeal 16S rRNA gene amplification, respectively. The amplifications was carried out under the following conditions: initial denaturation at 95°C for 5 min, followed by 30 cycles of denaturation at 95°C for 30 sec, primer annealing at 55°C for 30 sec, and extension at 72°C for 30 sec, followed by a final elongation at 72°C for 5 min. The amplified products were purified with the QIAquick PCR purification kit (Qiagen, Valencia, USA). Obtained reads from the different samples were sorted by the unique barcodes of each PCR product. The sequences of the barcode, linker, and primers were removed from the original sequencing reads. Potential chimera sequences were detected with Bellerophon, which involves comparing the BLASTN search results between the forward and reverse halfsequences [20]. Reads were assigned against the EzTaxon-e database (http://eztaxon-e. ezbiocloud.net) [21], which contains 16S rRNA gene sequences from type strains that have valid published names and representative species level phylotypes of either cultured or uncultured entries in the GenBank database with complete hierarchical taxonomic classification from the phylum to the species. The term uc means "unclassified taxon," and typical suffixes are _s (for species), _g (genus), _f (family), _o (order), _c (class) and _p (phylum) [22]. The results obtained from archaeal and bacterial communities, as well as from the different analysis methods, were compared with each other. From the pyrosequencing analysis, 15,043 bacterial sequence reads and 40,766 archaeal sequence reads were acquired. Low-quality and chimeric sequences were removed. The average read lengths for archaea and bacteria were 399 bp and 453 bp, respectively. To compare OTUs between samples, shared OTUs were obtained with the XOR analysis of the CLcommunity program (Chunlab Inc., Seoul, South Korea). Interactive Krona HTML5 [23] hierarchical and double pie chart community profiles have been included in the supplemental information online as charts_supplemental.zip (S1 File). Sequences from this study were deposited in the NCBI short-read archive under the accession number SRA051716.

Results and Discussion
Bacterial 16S rDNA PCR-DGGE . The band intensity of the major Sporomusa strain (S1-4) increased with the length of the fermentation period (Fig 2A). Sporomusa malonica is a Gram-negative spore-forming homoacetogen [24], and Sporomusa sphaeroides is a Gram-negative, sporeforming, banana-shaped bacteria with a described pH-range between 5.7 and 8.7 [25]. Sporomusa termida sp. nov., is an H 2 /CO 2 -utilizing acetogen isolated from termites [26], and Sporomusa paucivorans sp. nov., a methylotrophic bacterium that forms acetic acid from H 2 and CO 2 [27]. The Sporomusa sp. identified as a result of the DGGE band sequencing was found to be an acetogen. It was suggested that Sporomusa sp. could be acid-tolerant acetogens capable of activity in the pH range of 4.5-5.5. Additionally, under operating conditions of ORP = -430 mV [11], acetogens could be the dominant species in the bacterial community structure [28].

Archaeal 16S rDNA PCR-DGGE
The results of the archaeal 16S rDNA PCR-DGGE are shown in Fig 2C and 2D. Despite the analysis having been conducted only five days after inoculation with the anaerobic sludge, we found that the microbial community of the reactors was already established and dominant ( Fig  2C). At acidic pH, Methanosarcinaceae and Methanospirillum were dominant among the methanogens (Fig 2D). Methanospirillum is a hydrogenotrophic methanogen [29], and Methanosarcina is an acetoclastic methanogen [30]. Methanospirillum hungatei used formate or H 2 and CO 2 as substrates for CH 4 formation and growth [31]. Methanosaeta concilii was reported to use acetate and CO 2 as carbon sources [32]. The volume of produced CH 4 from 10 to 20 days was much more than that of the other periods (Fig 3), and the band M1-2, identified as Methanosaeta concillii, was detected with highest intensity in the lane of day 8 by DGGE. Therefore, both acetoclastic and hydrogenotrophic methanogen increased CH 4 production in days 10-15. As shown in Fig 2C and 2D, hydrogenotrophic and acetoclastic methanogens initially coexisted, but only the hydrogenotrophic methanogens were still alive at the end of the experiment. The stable presence of hydrogenotrophic methanogens indicates that a bubbleless membrane-diffusion device was successfully transferred the gaseous phase substrate. This device overcame the poor solubility of H 2 and afforded higher H 2 utilization efficiencies with consequent energy savings. Additionally, the CH 4 ratio of effluent gases reached 80-90% at day 7 and this ratio held until the end of operation [11,33]. This result showed that CH 4 production and microbial community structure became stable in the initial phase of operation.

Real-time qPCR of bacteria and archaea
To quantify the archaea with a role in CH 4 production, we used qPCR analysis with a specific primer to detect total archaea. Specifically, Methanobacteriales, representing hydrogenotrophic methanogens, and Methanomicrobiales, representing acetoclastic methanogens were tracked to discover their main mode of CH 4 production. Bacterial numbers were also quantified with general bacterial primers ( Table 3). The concentration of the total archaea ranged from 10 9 to 10 10 (copies/mL) during the operating periods of the reactors. All of the archaea, Methanobacteriales, and Methanomicrobiales had a similar pattern over time (Fig 3). Most of the archaea consisted of Methanobacteriales, but Methanomicrobiales were present at only levels of 0.01-1% (Fig  3). Methanobacteriales abundance dramatically increased during day 5, and the increased level was maintained afterwards. Although Methanomicrobiales abundance slightly increased temporarily from day 12 to 26, they returned to their initial concentration level over time (Fig 3).
The CH 4 production was highest at 15 h (Fig 3), due to synergistic effects from both the hydrogenotrophic and acetoclastic methanogens. Produced acetic acid was quickly consumed during this period, as shown in Fig 3. Acetic acid concentrations increased initially but quickly decreased, and no acetic acid was detected after day 36. It is believed that acetate was produced by acetogens, such as Sporomusa sp. (shown in DGGE band in Fig 2B) and it was consumed by acetoclastic methanogens like Methanosarcina (Fig 2D). Acetate consumption led to the dominance of hydrogenotrophic methanogens (Fig 2C). We suggest that the pH drop inhibited acetogen growth and inhibited acetate production. In a previous study, when another reactor was operated at neutral conditions (pH = 7), acetate concentration was stable at 4-6 g /L [11], whereas the only acetate consumption was in the acidic reactor. Additionally, the qPCR result showed a decrease of bacterial numbers with increasing fermentation time, indicating that low pH inhibited the growth of acetate-producing bacteria.
We also observed CH 4 consumption from day 14-21. Recently, methanogens were reported to have the ability to oxidize a small amount of CH 4 anaerobically when CH 4 was produced [34]. However, another study showed that the anaerobic oxidation of CH 4 was not directly mediated by methanogenic bacteria [35]. The reduction of sulfate or nitrate was related to CH 4 oxidation [36]. It is possible that the CH 4 oxidation level was high at~day 15 from sulfatereducing bacteria (the presence of nitrate reducing bacteria at ORP = -430 mV being very unlikely [11]) and decreased as the acidophilic Methanobacterium sp. became established in and then dominated [37] the microbial community.

Pyrosequencing of the microbial community
As shown in DGGE bands figure (Fig 2A and 2C), the DGGE did not show great microbial diversity and displayed only 2-4 bands. Some of the drawbacks of DGGE are the limited sequence information due to the small analyzed fragments (up to 500 bp) and its poor minor detection of small populations [38]. Additionally, it has been suggested that the DGGE for archaeal analysis should target each lower taxonomic groups [39]. Therefore, we used pyrosequencing for broad-based microbiome identification based on the sequencing-by-synthesis principle [40,41].
When the reactor had stabilized without pH adjustment, the microbial community was analyzed with pyrosequencing. Pyrosequencing was expected to show more detailed information and correlations between the archaeal and bacterial community. The results showed that 97.1% of archaeal sequence reads were assigned at the species level, indicating that the analysis results could provide sufficient resolution for archaeal community analysis. In contrast, while most of the bacteria sequence reads were assigned at the species level, a considerable proportion of sequence reads (16%) were assigned at the order level. Therefore, bacteria were analyzed at both the order and the species levels.
Pyrosequencing of the archaeal community. The taxonomic composition of methanogenic microbes is summarized in Table 4 (genus level) and Fig 4 (major classes (inner) and species (outer) in a double pie chart). The hydrogenotrophic methanogens Methanobacteriales dominated in the stabilized reactor. Methanobacteriales showed an abundance of 99% or more,  and the acetoclastic methanogen Methanosarcinales were detected at only 0.22% of the total portion of the microbial community. The pyrosequencing method provided results assigned at the species levels. As shown in Fig 4, three strains dominated among the species. Among them, AB236058 (uncultured Methanobacteriaceae) had the largest abundance at 49%. Methanobacterium congolense and M. subterraneum both also had an abundance greater than 10%. The strain AB236058 was detected in the CH 4 producing process and observed in low H 2 conditions [42]. Methanobacteriaceae became prevalent at low pH values [43]. Methanonacterium congolense is also a hydrogenotrophic methanogen. This strain uses CO 2 /H 2 for cell growth but not acetate [44]. Methanobacterium subterraneum is representative of CH 4 producing archaea and is capable of autotrophically growing in mineral medium without the addition of any organics [45]. Methanobacterium beijingense account for approximately 8% of the microbial community and has been reported to utilize H 2 /CO 2 and formate [46]. A microbial community producing CH 4 by hydrogenotrophy was developed without additional pH control. As previously stated, DGGE and real time PCR showed similar results with the pyrosequencing.
Pyrosequencing of the bacterial community. The taxonomic composition of the bacterial community is summarized in Table 5 (genus level) and Fig 5 (species level). Kim et al. reported that the methanogenic community shifted from acetoclastic methanogens to a hydrogenotrophic community, which is accompanied by an increase in the population of Firmicutes, in particular of Clostridia, in the bacterial community [47]. Thus, the stable bacterial community for CH 4 production was evaluated in terms of how it shifted. As shown in the Table 5, most bacterial sequence reads were assigned as uncultured strains at the genus level. Among them, AH009469 had an abundance of 20.58%. This strain belongs to Anaerolinaceae. Some strains of Anaerolinaceae are known to grow with hydrogenotrophic methanogens [33]. In addition, strains of Anaerolinaceae have been detected as a digester of CH 4 gas [48]. Next, AY214182 belonging to Spirochaetaceae had an abundance of 15.68% in the microbial community. Chartrain and Zeikus reported that Spirochaetaceae have been found to degrade ethanol to acetate in the presence of H 2 -consuming methanogens [49,50]. The strain assigned to EF198044 had an abundance of 8%. This strain was uncultured and detected in the mesophilic and thermophilic phenol-degrading methanogenic consortia [51]. At the class level, Clostridia showed an abundance of 17%. Clostridia are known as strong H 2 producers [52]. In addition, Clostridia can produce organic acids such as acetic acid by degrading organic substances. These characteristics of the microbial community would affect the methanogenic consortia. Both DGGE and pyrosequencing analysis showed the enrichment of hydrogenotrophic methanogen in Hf-MBfR without pH control. The time profile of DGGE bands did not show substantial diversity in the microbial community, but pyrosequencing indicated the enrichment of hydrogenotrophic methanogens in the whole microbial community even though it supplied less information about bacterial community at the species level. Additionally, qPCR data evinced both the enrichment of hydrogenotrophic methanogens and the pH inhibition on bacterial growth. These analyses were helpful in understanding the relationship between the bacteria and the methanogens in the Hf-MBfR (Fig 6). The acidic methanogenic reactor showed  The microbial community development in acidic methanogenic reactor. The acetate was produced by acetogens from H 2 and CO 2 . Acetate was used by bacteria and H 2 and CO 2 were regenerated. Finally, acetate was limited, but H 2 continued to be available, and hydrogenotrophic methanogens were harvested.
doi:10.1371/journal.pone.0144999.g006 hydrogenotrophic methanogen enrichment, and low pH seems to have inhibited acetogen growth. Limited acetate production led to the growth of hydrogenotrophic methanogens rather than acetoclastic methanogens. Low pH inhibited bacterial growth, especially of acetogens as acetate suppliers to acetoclastic methanogens, and enriched hydrogenotrophic methanogens (Fig 6).

Conclusions
In this study, we investigated the characteristics of the microbial community in a hollow-fiber membrane biofilm reactor (Hf-MBfR), focusing on the presence of hydrogenotrophic methanogens that convert supplied CO 2 and H 2 into CH 4. The reactor was operated under acidic conditions, and the microbial community was analyzed with 16S rDNA PCR-DGGE, real-time qPCR, and pyrosequencing. The results showed a stable microbial community was established relatively early in the fermentation, and this community was enriched with hydrogenotrophic methanogens (99.6% of archaea). Therefore, the high conversion efficiency of CO 2 to CH 4 was induced by the enrichment of hydrogenotrophic methanogens by the acidic operation of the Hf-MBfR, which mitigated the disadvantage of H 2 as electron donors (less soluble in water, explosive gas easily released from air diffuser). The community structure showed that the Hf-MBfR properly supplied the hydrogen for hydrogenotrophic methanogen. Additionally, the result suggested that the acidic operation of Hf-MBfR inhibited acetogens and led to the enrichment of hydrogenotrophic methanogens, achieving a high conversion ratio of CO 2 to CH 4 .