Microbial Consortiums of Hydrogenotrophic Methanogenic Mixed Cultures in Lab-Scale Ex-Situ Biogas Upgrading Systems under Different Conditions of Temperature, pH and CO

In this study, hydrogenotrophic methanogenic mixed cultures taken from 13 lab-scale ex-situ biogas upgrading systems under different temperature (20–70 °C), pH (6.0–8.5), and CO (0–10%, v/v) variables were systematically investigated. High-throughput 16S rRNA gene sequencing was used to identify the microbial consortia, and statistical analyses were conducted to reveal the microbial diversity, the core functional microbes, and their correlative relationships with tested variables. Overall, bacterial community was more complex than the archaea community in all mixed cultures. Hydrogenotrophic methanogens Methanothermobacter, Methanobacterium, and Methanomassiliicoccus, and putative syntrophic acetate-oxidizing bacterium Coprothermobacter and Caldanaerobacter were found to predominate, but the core functional microbes varied under different conditions. Multivariable sensitivity analysis indicated that temperature (p < 0.01) was the crucial variable to determine the microbial consortium structures in hydrogenotrophic methanogenic mixed cultures. pH (0.01 < p < 0.05) significantly interfered with the relative abundance of dominant archaea. Although CO did not affect community (p > 0.1), some potential CO-utilizing syntrophic metabolisms might be enhanced. Understanding of microbial consortia in the hydrogenotrophic methanogenic mixed cultures related to environmental variables was a great advance to reveal the microbial ecology in microbial biogas upgrading process.


Introduction
Biogas upgrading via an external H 2 supply to promote CO 2 biomethanation has attracted considerable attention [1]. Hydrogenotrophic methanogens play an important role in the efficient biogas upgrading process and can capture CO 2 by combining it with H 2 to produce CH 4 [2,3]. It was reported that the methane content in biogas could be increased up to 89-96% by hydrogenotrophic methanogens in various in-situ and ex-situ CO 2 biomethanation processes [4][5][6][7][8]. Microbial analysis conducted in various biogas upgrading configurations revealed that the Methanobacterium, Methanothermobacter, Methanoculleus, and Methanomicrobium species were dominant in the hydrogenotrophic methanogenic consortium [3,5,[9][10][11][12]. The utilization of hydrogenotrophic methanogenic mixed cultures enriched

DNA Extraction
In this study, total community genomic DNA was extracted using an E.Z.N.A. ® Mag-Bind Soil DNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer's instructions. The DNA concentration was measured using a Qubit 2.0 fluorimeter (Life Technologies, Waltham, MA, USA), and the quality of the DNA was checked using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara" CA, USA) to ensure that adequate amounts of high-quality genomic DNA were extracted.

Polymerase Chain Reaction
Polymerase chain reaction (PCR) was conducted immediately after DNA extraction. The 16S rRNA V3-V4 gene was amplified using KAPA HiFi Hot Start ready mix (TaKaRa Bio Inc., Shiga, Japan). Archaeal DNA was amplified using the forward primer 340F (CCCTAYGGGGYGCASCAG) and the reverse primer 1000R (GGCCATGCACYWCYTCTC) to generate the first-round amplicon and the primers 349F (GYGCASCAGKCGMGAAW) and 806R (GGACTACVSGGGTATCTAAT) to generate the second-round amplicon [23] with the reactions set up as follows: 2 µL of microbial DNA (10 ng/µL), 1 µL each of the forward and reverse primers (10 µM), and 15 µL of 2 × KAPA HiFi Hot Start Ready Mix (total volume of 30 µL). The plate was sealed and PCR was performed using an Applied Biosystems 9700 instrument (Thermo Fisher Scientific, MA, USA) and the following program: 1 cycle of denaturation at 95 • C for 3 min; 5 cycles of denaturation at 95 • C for 30 s, annealing at 45 • C for 30 s, and elongation at 72 • C for 30 s; followed by 20 cycles of denaturation at 95 • C for 30 s, annealing at 55 • C for 30 s, and elongation at 72 • C for 30 s; with a final extension at 72 • C for 5 min. Bacterial DNA was amplified using the PCR primers forward 341F (CCTACGGGNGGCWGCAG) and reverse 805R (GACTACHVGGGTATCTAATCC), with the reaction set up as described for the archaeal PCR [23]. The PCR products were assessed via electrophoresis in 1% (w/v) agarose gels in TBE buffer (Tris, boric acid, and EDTA) stained with ethidium bromide (EB) and visualized under UV light.

DNA Purification and Quantitative Mixing
AMPure XP beads (Beckman Coulter, USA) were used to remove the free primers and primer dimer species from the amplicon products. A library was constructed using the universal Illumina adaptor and index. Depending on the coverage needs, all libraries were pooled for one run. Amplicons from each reaction mixture were pooled in equimolar ratios based on their concentration. The reaction products were sequenced using the Illumina MiSeq system (Illumina, San Diego, CA, USA) according to the manufacturer's instructions. After sequencing, data were collected as follows: (1) Two short Illumina readings were assembled by PEAR (v0.9.6) software according to the overlap, and fastq files were processed to generate individual fasta and qual files, which could then be analyzed by standard methods; (2) Sequences containing ambiguous bases as well as any longer than 480 base pairs (bp) were dislodged and those with a maximum homopolymer length of 6 bp were allowed [24]. In addition, sequences shorter than 200 bp were removed; (3) All identical sequences were merged into one; (4) Sequences were aligned according to a customized reference database; (5) The intactness of the index and the adaptor was checked, and all sequence information corresponding to the index and adaptor sequences was then removed; (6) Noise was removed using the Pre.cluster tool. Chimeras were detected using Chimera UCHIME. All the software was in the Mothur package [25].

Statistical Data Analysis
All the sequences were grouped into operational taxonomic units (OTUs) using a program from the Mothur package [26]. Effective sequences after processing were submitted to the Ribosomal Database Project (RDP) Classifier to confirm the archaeal and bacterial sequences.

Alpha Diversity Analysis
The microbial diversity of each single sample was estimated by alpha diversity analysis. In this study, species richness and diversity statistics, including the Coverage, Chao1, ACE, Simpson, and Shannon indices, were calculated according to the OTU results. Then, all the effective sequences without primers were submitted for downstream analysis [27].

Microbial Community Analysis
The microbial community was illustrated using a heatmap, presenting the composition of microorganisms and their relative abundances. Thereafter, the core functional archaea and bacteria in the hydrogenotrophic methanogenic mixed culture were analyzed using network diagrams calculated using the Kamada-Kawai algorithm with R programming language and the Vegan package.

Correlative Analysis Between Microbial Community and Environmental Variables
Correlation coefficients of single microorganisms to each environmental variable were evaluated by Pearson's correlation algorithm. Eventually, the multivariable sensitivity of the overall hydrogenotrophic methanogenic mixed cultures was revealed by redundancy analysis (RDA) [28] and canonical correlation analysis (CCA) [29] with R programming language and the Vegan package. The selection of RDA or CCA was determined by gradient lengths of the detrended correspondence analysis. If the value of its first axis was lower than 3.5, CCA was used, otherwise the RDA was more appropriate.

Overview of Community Diversity under Various Temperature, pH, and CO Conditions
Microbial DNA sequences were generated through Miseq from the hydrogenotrophic methanogenic mixed cultures. After quality control, a total of 527,952 archaeal sequences and 560,780 bacterial sequences were obtained. These sequences were subsequently clustered into OTUs at 97% sequence similarity. With the purpose of the assessing the internal complexity of microbial community, index of Coverage, Chao, ACE, Shannon and Simpson indices were investigated. A coverage estimator > 0.999 indicates that the majority of bacterial and archaeal OTUs were captured, guaranteeing the reliability of the sequencing results ( Table 2). As shown in Table 2, the number of OTUs, as well as values for the richness and evenness indices of the bacterial communities, were significantly higher than those of the archaea community, demonstrating that the bacterial community was more complex than the archaea community in the hydrogenotrophic methanogenic mixed culture. In the archaeal community, values for both Chao and ACE indices were lower than 16 in the 55B, 70B, and 70N_10 mixed cultures, indicating that the corresponding conditions presented higher selectivity towards archaea diversity. By contrast, for the 20N, 30N, 55N_5, 70N, and 70A samples, the conditions were relatively beneficial for the richness of the archaea community, with observed Chao and ACE index values greater than 24. Further comparisons of the Shannon and Simpson indices for different mixed cultures revealed that there was regularity behind the variety of indices. The evenness deteriorated with increased temperature in the archaeal community, represented by lower Shannon but higher Simpson values. Non-neutral pH negatively influenced the archaeal community, and the values for the Simpson index, for example, followed the order 55A > 55B > 55N and 70A > 70B > 70N, suggesting that the majority of archaea survived under neutral conditions (pH of 7.0). It was reported that the decreased archaeal diversity under alkaline conditions could have contributed to the greater energy consumption required to sustain the intracellular pH balance [30]. In this study, acidic conditions were found to severely inhibit archaeal community diversity. This might result from the presence of acid-sensitive archaea in the hydrogenotrophic methanogenic mixed cultures.
As for the bacterial community, the corresponding conditions for the 20N and 30N samples were more favorable than for the others, with both Chao and ACE indices higher than 290 (Table 2). Overall, the richness of the bacterial community decreased with the gradual increase in temperature, with lower values for Chao and ACE indices obtained at 55 and 70 • C, which a similar trend for the archaeal community. Notably, the growth of bacterial communities remained relatively unstimulated at non-neutral pH. It is interesting that bacterial diversity in the alkaline hydrogenotrophic methanogenic mixed culture was enhanced under extreme-thermophilic conditions (for example, a Chao index of 110.11 for 70N and 223.84 for 70B). This result might due to the fact that the dominant bacteria consortium in the mixed culture was susceptible to alkaline conditions, such that the relative abundances of other bacteria were increased as a result of their higher alkaline tolerance. Further evidence is presented in the following sections.
With the exception of temperature and pH variables, the microbial characteristics related to CO introduction in terms of OTUs as well as Coverage, richness, and evenness indices exhibited irregular fluctuations under various temperature and pH conditions. Therefore, the microbial community composition and structure in different hydrogenotrophic methanogenic mixed cultures should be further investigated.

Microbial Community Structure of Hydrogenotrophic Methanogenic Mixed Cultures
Taxonomic classification at the genus level was systematically examined to verify the structure of microbial communities. Overall, 8 archaeal genera and 34 bacterial genera (relative abundance > 3%) were identified and are illustrated in Figure 1 (the relative abundance of all detected microbial genera in the mixed cultures are provided in Tables S3 and S4). Figure 1a shows the structures of the archaeal communities at the genus level in a heatmap. As can be seen, hydrogenotrophic methanogens dominated in all samples with a narrow distribution, mainly including Methanothermobacter, Methanobacterium, and Methanomassiliicoccus. The above three genera accounted for 79.2-99.8% of archaeal abundance in total. By contrast, the typical acetoclastic archaea Methanosaeta showed inferior growth in all hydrogenotrophic methanogenic mixed cultures with a relative abundance of less than 9.5%. The other genera were marginally detected in partial samples due to their competitive disadvantage. The unclassified/no-rank microbial genera in the mixed cultures are marked with their upstream taxonomic name; for example, the no-rank archaea at the genus level within phylum Bathyarchaeota is described as norank p Bathyarchaeota.
As illustrated from Figure 1a, the sequences related to the genera Methanobacterium (67.6% and 65.4% in the 20N and 30N samples, respectively) and Methanomassiliicoccus (18.6% and 20.6% in the 20N and 30N samples, respectively) showed a relatively high abundance in the above-mixed cultures. With the temperature increasing to 55, 65, and 70 °C , the genus Methanothermobacter demonstrated dominance over Methanobacterium and Methanomassiliicoccus with a relative abundance of 38.0-97.7%, due to its favorable growth at higher temperatures [31]. This finding was in close agreement with a previous study wherein Methanothermobacter dominated the archaeal communities in ex-situ methanation systems at 55 and 65 °C [8]. Interestingly, pH was a significant factor of microbial abundance perturbation that contributed to a 22.1-30.1% decrease in Methanothermobacter, but a 23.31-50.8% increase in Methanobacterium under alkaline conditions. A pH of 8.5 seemed to greatly The unclassified/no-rank microbial genera in the mixed cultures are marked with their upstream taxonomic name; for example, the no-rank archaea at the genus level within phylum Bathyarchaeota is described as norank p Bathyarchaeota.
As illustrated from Figure 1a, the sequences related to the genera Methanobacterium (67.6% and 65.4% in the 20N and 30N samples, respectively) and Methanomassiliicoccus (18.6% and 20.6% in the 20N and 30N samples, respectively) showed a relatively high abundance in the above-mixed cultures. With the temperature increasing to 55, 65, and 70 • C, the genus Methanothermobacter demonstrated dominance over Methanobacterium and Methanomassiliicoccus with a relative abundance of 38.0-97.7%, due to its favorable growth at higher temperatures [31]. This finding was in close agreement with a previous study wherein Methanothermobacter dominated the archaeal communities in ex-situ methanation systems at 55 and 65 • C [8]. Interestingly, pH was a significant factor of microbial abundance perturbation that contributed to a 22.1-30.1% decrease in Methanothermobacter, but a 23.31-50.8% increase in Methanobacterium under alkaline conditions. A pH of 8.5 seemed to greatly alleviate the negative effects of the extreme-thermophilic condition for Methanobacterium. The higher alkaline adaptability of the genus Methanobacterium was consistent with the results of a site remediation study wherein Methanobacterium sp. dominated in a lime kiln filtrate [32]. Additionally, it is noteworthy that the introduction of CO may not severely interfere with the dominant archaeal genera, as it has been reported that Methanothermobacter spp. are capable of converting CO to CH 4 [33,34]. The relative abundance of Methanothermobacter in the samples where CO was introduced remained at 35.1-41.3% and 99.0-99.8% at 50 and 70 • C, respectively.
In terms of bacterial relative abundance, significant dominance was observed with two bacterial phyla, namely Firmicutes and Proteobacteria (61.2-88.0%). As illustrated in Figure 1b, the genera Thioclava and Sulfurovum were the most abundant at lower temperatures (a total of 47.8 and 36.9% in the 20N and 30N samples, respectively), followed by Proteocatella (6.3-14.1%), Sulfuricurvum (1.8-12.1%), norank family Anaerolineaceae (3.2-11.0%), Proteiniclasticum (5.7-9.1%), Sedimentibacter (2.9-4.2%) and Longilinea (1.8-3.5%). By comparison, a remarkable spatial difference was demonstrated whereby the genus Coprothermobacter dominated in thermophilic mixed cultures with a relative abundance of 42.8 to 59.0%, and Caldanaerobacter abundance was prominently increased, accounting for 10.9-49.2% in extreme-thermophilic mixed cultures. These observations are all in good accordance with their reported optimal growth temperatures [35,36]. Notably, Coprothermobacter was reported as a typical syntrophic bacterial genus that contributes to the degradation of small-molecule organic compounds to promote hydrogen production, subsequently providing sufficient raw materials for hydrogenotrophic methanogens [37,38]. Moreover, the genus Coprothermobacter was also a potential syntrophic acetate oxidizer interacting with methanogens at 55 • C, as previous reported [39]. Hence, it might be involved in converting acetate to H 2 and CO 2 , thus providing hydrogenotrophic methanogens with a slight amount of carbon for further biomethanation. Additionally, Caldanaerobacter, which dominated at 70 • C, was also a potential syntrophic acetate oxidation bacterium (SAOB) that established an intimate syntrophic association with extreme-thermophilic methanogens in the mixed cultures [40]. It is interesting that four abundant bacterial genera Tepidiphilus, Exiguobacterium, norank order D8A-2, and Tepidanaerobacter, where were subdominant in the thermophilic and extreme-thermophilic samples, were also reported as potential SAOBs [41][42][43][44]. These syntrophic genera, together with Coprothermobacter and Caldanaerobacter, accounted for the majority of bacterial abundance, ranging from 39.9 to 86.0%, implying the pathway of syntrophic acetate oxidation coupled with hydrogenotrophic methanogenesis (SAO-HM) existed in the hydrogenotrophic methanogenic mixed culture. Moreover, the total proportion of SAOBs that dominated in the samples followed the sequence of 55N (83.1%) > 55B (79.5%) > 70N (50.8%) > 70B (41.7%), indicating the pathway of SAO-HM might be weakened at 70 • C, especially under alkaline conditions. Besides, CO introduction seemed to interfere with the bacterial proportions by augmenting Coprothermobacter abundance by 5.5-16.3% but significantly decreasing Caldanaerobacter abundance by 34.8-36.2%. Coincidentally, in a previous investigation on simultaneous sewage sludge treatments and CO biomethanation, an increased Coprothermobacter abundance was also observed in the liquid phase after CO addition [25]. It is noteworthy that although some species belonging to the genus Caldanaerobacter have been reported to utilize CO as a substrate [45], they may suffer a comparative competitive disadvantage in hydrogenotrophic methanogenic mixed cultures. Overall, the bacterial communities in hydrogenotrophic methanogenic mixed cultures were abundant with SAOBs, of course, as well as a small amount of various indispensable bacteria.

Correlative Relationship Between Microorganisms and Temperature, pH, and CO
The correlative relationship between the environmental variables (temperature, pH, and CO) and the relative abundance of microbial genera in the hydrogenotrophic methanogenic mixed cultures were examined using Pearson correlation analysis. Archaeal and bacterial genera that were significantly correlated with at least one of the environmental variables are summarized in Table 3. The overall archaeal community dissimilarities were primarily attributed to the temperature and pH variables, to which they were negatively correlated. Specifically, Methanobacterium (r = −0.893, p < 0.01), Methanomassiliicoccus (r = −0.932, p < 0.01), norank phylum Bathyarchaeota (r = −0.823, p < 0.01), norank phylum ARC26 (r = −0.812, p < 0.01), and norank family Terrestrial Miscellaneous Gp TMEG (r = −0.603, 0.05 < p < 0.1) demonstrated a negative relationship with temperature, while Methanothermobacter (r = −0.732, 0.01 < p < 0.05), Methanosarcina (r = −0.584, 0.05 < p < 0.1), Methanobrevibacter (r = −0.725, 0.01 < p < 0.05), Methanoculleus (r = −0.824, p < 0.01), Methanospirillum (r = −0.650, p < 0.1), Methanosphaera (r = −0.568, p < 0.1), and norank family Thermoplasmatales Incertae Sedis (r = −0.826, p < 0.01) exhibited a significant negative correlative relationship with pH. The megatrends of archaeal abundance decline related to increasing temperature and pH variables were highly consistent to the alpha diversity results in Section 3.1, indicating high temperature and pH were unfavorable to the majority of methanogens in hydrogenotrophic methanogenic mixed cultures. However, a significant positive correlation coefficient was observed between the genus Methanothermobacter and temperature (r = −0.560, 0.05 < p < 0.01), explaining the predominance of Methanothermobacter over all the archaeal genera in the extreme-thermophilic mixed cultures. In terms of CO introduction, in general, the archaeal abundance was insignificantly influenced. Although the genus Methanomethylovorans showed a positive significant correlation to increasing CO concentration (r = 0.603, p < 0.1), the maximum relative abundance of 3.5% was negligible compared to the dominant Methanothermobacter or Methanobacterium.
Dynamic changes in the bacterial community as a result of temperature, pH, and CO variables exhibited a higher complexity than for archaea. As observed in Table 3, a total of 19 bacterial genera were significantly influenced by temperature, among which, 13 genera were positively enriched but 6 genera were negatively inhibited. By contrast, higher pH and CO introduction seemed to be favorable for bacterial communities, with 6 and 3 positively correlative genera, respectively. Potential syntrophic metabolism was implied by the enhancement of abundance for members of the unclassified family Thermoanaerobacteraceae in response to CO introduction (r = 0.753, 0.0 5 < p < 0.01). The family Thermoanaerobacteraceae has been previously reported to be capable of utilizing CO. Alves et al. [46] suggested that the genus Thermoanaerobacter belonging to family Thermoanaerobacteraceae showed high resistance to CO-containing environments. This contributed to the conversion of CO into direct materials for hydrogenotrophic methanogenesis or acetotrophic methanogenesis. For instance, Thermoanaerobacter thermohydrosulfuricus subsp. carboxydovorans (stain TDL) was able to convert CO/H 2 O to CO 2 /H 2 [45], and Thermoanaerobacter kivui was capable of capturing CO as a single-electron source for CH 3 COOH/H 2 formation and coupling with energy conservation for oxidation [47]. In this study, genus unclassified family Thermoanaerobacteraceae accounted for 4.0% and 21.3% of total bacterial abundance in 10% CO-containing samples at 55 and 70 • C, respectively, while it was fairly indetectable in non-CO mixed cultures. These results strongly suggest the possibility of syntrophic CO-utilization associated with methanogens in the hydrogenotrophic methanogen mixed cultures. Additionally, CO acclimation at 70 • C increased the relative abundance of Dictyoglomus (from 6.4 to 8.3%), Brockia (from 1.1 to 6.3%) and Thermanaeromonas (from 0.2 to 8.1%), which were identified as sulfur-reducing bacteria [48][49][50]. It was likely that the supplied CO promoted electron transfer for the sulfur reduction. However, the relative abundance of the three abovementioned sulfur-reducing bacteria reduced to 1.8%, 2.8%, and 1.2% after 10% CO exposure, indicating that a higher CO concentration might not be favorable for sulfur-reducing syntrophic associations. Conversely, the abundance of the genus Coprothermobacter increased both at 55 • C (from 41.3 to 50.3%) and 70 • C (from 0.3 to 15.3%), implying that the SAO-HM pathway established by Coprothermobacter was enhanced by CO [21].
Furthermore, a profile network analysis was constructed to identify the core functional microorganisms in the hydrogenotrophic methanogenic mixed cultures. After qualified clustering, 10 effective archaeal nodes (relative abundance > 0.1%) and 49 effective bacterial nodes (relative abundance > 1%) related to hydrogenotrophic methanogenic mixed cultures were identified and are illustrated in Figure 2. The genus Methanothermobacter was observed as the core functional archaeal microbe (degree of 11) in the majority of hydrogenotrophic methanogenic mixed cultures. This result is highly consistent with the previous observation that the genus Methanothermobacter is responsible for the majority of biological hydrogen methanation [51]. However, it was replaced by the sub-core functional genus Methanobacterium (degree of 6) in the 20N and 30N samples due to an unfavorable environmental temperature of below 40 • C [52]. Regarding bacterial microbes, Coprothermobacter was identified as the core functional bacterial microbe (degree of 9) despite it being undetected in 20N and 30N. It is noteworthy that the sub-core functional bacterial genus Caldanaerobacter (degree of 9) was also inexistent at 20 and 30 • C. Conversely, it was dominated by the genus Thioclava (degree of 2). Actually, more than half of the bacterial nodes connected to 20N and 30N were unidentified under thermophilic or extreme-thermophilic conditions. Overall, this demonstrates that the growth and presence of the core functional bacterial microbes are very much determined by environmental temperature variables.
is highly consistent with the previous observation that the genus Methanothermobacter is responsible for the majority of biological hydrogen methanation [51]. However, it was replaced by the sub-core functional genus Methanobacterium (degree of 6) in the 20N and 30N samples due to an unfavorable environmental temperature of below 40 °C [52]. Regarding bacterial microbes, Coprothermobacter was identified as the core functional bacterial microbe (degree of 9) despite it being undetected in 20N and 30N. It is noteworthy that the sub-core functional bacterial genus Caldanaerobacter (degree of 9) was also inexistent at 20 and 30 °C . Conversely, it was dominated by the genus Thioclava (degree of 2). Actually, more than half of the bacterial nodes connected to 20N and 30N were unidentified under thermophilic or extreme-thermophilic conditions. Overall, this demonstrates that the growth and presence of the core functional bacterial microbes are very much determined by environmental temperature variables.

The Multivariable Sensitivity of Hydrogenotrophic Methanogenic Mixed Cultures
RDA and CCA were further performed to assess the contributions of environmental variables (temperature, pH, and CO) to variances in the archaeal and bacterial communities. As illustrated in Figure 3a,b, the length of the red arrows indicates the extent to which the microbial community characteristics were influenced by the environmental variables. For archaeal community, the relative influence of the environmental variables followed the sequence of temperature (r 2 = 0.8658, p < 0.01), pH (r 2 = 0.6335, 0.01 < p < 0.05), and CO (r 2 = 0.1546, p > 0.1). This suggests that temperature and pH significantly determine the characteristics of the archaeal community in hydrogenotrophic methanogenic mixed cultures. Interestingly, a similar multivariable sensitivity was observed in the bacterial community where the temperature (r 2 = 0.9524, p < 0.01) and pH (r 2 = 0.8143, 0.01< p < 0.05) significantly influenced the bacterial structure in the hydrogenotrophic methanogenic mixed cultures. Although the insignificant correlation of CO introduction (r 2 = 0.3106, p > 0.1) was maintained, the increasing r 2 (from 0.1546 to 0.3106) implies an enhancement of the potential syntrophic associations in CO utilization. RDA and CCA were further performed to assess the contributions of environmental variables (temperature, pH, and CO) to variances in the archaeal and bacterial communities. As illustrated in Figure 3a,b, the length of the red arrows indicates the extent to which the microbial community characteristics were influenced by the environmental variables. For archaeal community, the relative influence of the environmental variables followed the sequence of temperature (r 2 = 0.8658, p < 0.01), pH (r 2 = 0.6335, 0.01 < p < 0.05), and CO (r 2 = 0.1546, p > 0.1). This suggests that temperature and pH significantly determine the characteristics of the archaeal community in hydrogenotrophic methanogenic mixed cultures. Interestingly, a similar multivariable sensitivity was observed in the bacterial community where the temperature (r 2 = 0.9524, p < 0.01) and pH (r 2 = 0.8143, 0.01< p < 0.05) significantly influenced the bacterial structure in the hydrogenotrophic methanogenic mixed cultures. Although the insignificant correlation of CO introduction (r 2 = 0.3106, p > 0.1) was maintained, the increasing r 2 (from 0.1546 to 0.3106) implies an enhancement of the potential syntrophic associations in CO utilization. Additionally, the co-influences of multivariable on hydrogenotrophic methanogenic mixed cultures are implied by the angle between the environmental variable arrows. Specifically, angle < 90° represents a positive relationship, while angle > 90° indicates a negative relationship. In Figure  3a, the angles between the three variables demonstrate that a positive relationship existed between temperature, pH, and CO co-influence on archaeal community. This suggests that the introduction of any two variables contributed, to a certain extent, to promoting the changing tendency of the overall archaeal community caused by the third variable. By contrast, a negative co-influence was observed between pH and CO in the bacterial community when angle > 90° (Figure 3b). This implies a mutually restricted relationship between the influence of the pH and CO variables on the overall bacterial community. After comparison, the bacterial community in hydrogenotrophic methanogenic mixed cultures showed higher multivariable sensitivity than the archaeal community regarding temperature, pH, and CO.
It is noteworthy that the angle between the environmental variables and functional microbes confirmed the correlative relationship observed in the above section. For instance, an angle between the pH variable and the core functional archaeal genus Methanothermobacter of >90° demonstrated the negative influence of pH on the relative abundance of Methanothermobacter, which was highly consistent with the Pearson correlation coefficients (r = −0.732, 0.01 < p < 0.05) in Section 3.3. Similarly, the positive influence of the temperature variable on the genus Methanothermobacter was confirmed Additionally, the co-influences of multivariable on hydrogenotrophic methanogenic mixed cultures are implied by the angle between the environmental variable arrows. Specifically, angle < 90 • represents a positive relationship, while angle > 90 • indicates a negative relationship. In Figure 3a, the angles between the three variables demonstrate that a positive relationship existed between temperature, pH, and CO co-influence on archaeal community. This suggests that the introduction of any two variables contributed, to a certain extent, to promoting the changing tendency of the overall archaeal community caused by the third variable. By contrast, a negative co-influence was observed between pH and CO in the bacterial community when angle > 90 • (Figure 3b). This implies a mutually restricted relationship between the influence of the pH and CO variables on the overall bacterial community. After comparison, the bacterial community in hydrogenotrophic methanogenic mixed cultures showed higher multivariable sensitivity than the archaeal community regarding temperature, pH, and CO.
It is noteworthy that the angle between the environmental variables and functional microbes confirmed the correlative relationship observed in the above section. For instance, an angle between the pH variable and the core functional archaeal genus Methanothermobacter of >90 • demonstrated the negative influence of pH on the relative abundance of Methanothermobacter, which was highly consistent with the Pearson correlation coefficients (r = −0.732, 0.01 < p < 0.05) in Section 3.3. Similarly, the positive influence of the temperature variable on the genus Methanothermobacter was confirmed by the angle between the temperature arrow and Methanothermobacter (<90 • ), which also closely coincided with the correlation analysis (r = 0.560, 0.05 < p < 0.1). For the same reason, the dependence of the sub-core functional archaeal genera on variables of temperature, pH, and CO was clearly illustrated and exhibited identical results as before. The same relationship results as were mentioned in Section 3.3 were observed in the bacterial community (Figure 3b).

Conclusions
Hydrogenotrophic methanogenic consortia, with hydrogenotrophic methanogens Methanothermobacter, Methanobacterium and Methanomassiliicoccus, and putative SAOB Coprothermobacter and Caldanaerobacter were eventually constructed in lab-scale ex-situ biogas upgrading systems. Temperature and pH significantly affected the microbial community. Dominant archaea and bacteria shifted from Methanobacterium to Methanothermobacter, and Coprothermobacter to Caldanaerobacter, respectively, with increasing temperature from 20 to 70 • C. Notably, Methanothermobacter deteriorated but Methanobacterium was augmented under alkaline conditions at high temperature. While CO introduction did not change the dominance of Methanothermobacter in mixed cultures, but some potential syntrophic association was probably enhanced. Tested environmental variables affected the archaea structure followed the sequence of temperature, pH, and CO.