Novel Long-Chain Fatty Acid (LCFA)-Degrading Bacteria and Pathways in Anaerobic Digestion Promoted by Hydrochar as Revealed by Genome-Centric Metatranscriptomics Analysis

ABSTRACT

A large amount of long-chain fatty acids (LCFA) are generated after lipids hydrolysis in anaerobic digestion (AD), and LCFA are difficult to be biodegraded. This study showed that hydrochar (HC), which was produced during the hydrothermal liquefaction of organic wastes, significantly increased the methane production rate (by 56.9%) of oleate, a typical refractory model LCFA. Genomic-centric metatranscriptomics analysis revealed that three novel microbes (Bin138 Spirochaetota sp., Bin35 Smithellaceae sp., and Bin54 Desulfomonilia sp.) that were capable of degrading LCFA were enriched by HC, which played an important role in the degradation of oleate. LCFA was degraded to acetate through the well-known LCFA β-oxidation pathway and the combined β-oxidation and butyrate oxidation pathway. In addition, it was found that HC promoted the direct interspecies electron transfer (DIET) between Methanothrix sp. and Bin54 Desulfomonilia sp. The enriched new types of LCFA-degrading bacteria and the promotion of DIET contributed to the improved methane production rate of oleate by HC.

IMPORTANCE Long-chain fatty acids (LCFA) are difficult to be degraded in anaerobic digestion (AD), and the known LCFA degrading bacteria are only limited to the families Syntrophomonadaceae and Syntrophaceae. Here, we found that hydrochar effectively promoted AD of LCFA, and the new LCFA-degrading bacteria and a new metabolic pathway were also revealed based on genomic-centric metatranscriptomic analysis. This study provided a new method for enhancing the AD of organic wastes with high content of LCFA and increased the understanding of the microbes and their metabolic pathways involved in AD of LCFA.

INTRODUCTION

Anaerobic digestion (AD) is currently one of the most effective methods for the treatment of organic wastes, which can produce renewable energy methane and at the same time reduce the amount of waste (1). Organic wastes are mainly composed of carbohydrates, proteins, and lipids. Compared with carbohydrates and proteins, lipids possess the highest methanogenic potential of 1.014 LCH₄/gVS (2). However, if the lipid content in the AD system is too high, it easily becomes a bottleneck step for AD (3). This is mainly because the hydrolysis of lipids generates a large amount of glycerol and long-chain fatty acids (LCFA). LCFA are easily adsorbed on the surface of microbes, forming a barrier layer that hinders the absorption of nutrients (4). Moreover, the accumulation of LCFA on the surface of cell membranes can also prevent microbes from regulating homeostasis, such as intracellular pH (5). In addition, LCFA needs to undergo repeated cycles of β-oxidation degradation. Each β-oxidation process removes two carbon atoms and produces acetate and H₂. The oxidation of LCFA is an endergonic reaction under standard conditions and requires a low H₂ partial pressure to ensure stable operation (6). Therefore, hydrogenotrophic methanogens need to consume excess H₂ through interspecies hydrogen transfer (IHT), maintaining the reaction thermodynamically favorable (7). Ensuring the fast degradation of LCFA is of great importance for the treatment of organic wastes with high lipids content.

To alleviate the inhibitory stress of LCFA on AD and promote its degradation, many studies chose various materials as additives to stimulate the AD of LCFA, such as zeolite, granular activated carbon (GAC), biochar, etc. (8–10). The addition of adsorptive materials such as zeolite and bentonite to alleviate the inhibition of LCFA has been widely used, mainly by reducing the concentration of LCFA in the system to alleviate the toxicity to microbes (8, 11). On the other hand, the addition of materials such as biochar, carbon nanotubes, and GAC could also promote the degradation of LCFA in the AD system (10, 12). It was thought to be due to the establishment of direct interspecies electron transfer (DIET) between syntrophic bacteria and methanogenic archaea, although there was generally little evidence. DIET can achieve electrons exchange through biological electrical connections, which is a more energy-conserving process and a feasible alternative to IHT in terms of thermodynamics, increasing the efficiency of methanogenesis (7, 13). In recent studies, the application of hydrochar (HC) has attracted attention, which effectively enhanced the methane production rate during the AD (14–17). HC is the solid product obtained by hydrothermal liquefaction of biomass, which is a green by-product of organic wastes and has a higher energy density compared with raw biomass (18). This popular material can be used to adsorb heavy metal ions, organic pollutants, or other substances (19, 20) and promote DIET through surface oxygen-containing functional groups (17). However, the influence of HC on the degradation of LCFA has not been studied, and it is worth investigating how HC affects AD of LCFA.

At present, the LCFA-degrading syntrophic bacteria proved by pure culture or in coculture with hydrogen-consuming microbes belong to the families Syntrophomonadaceae and Syntrophaceae, within the phyla Firmicutes and Deltaproteobacteria, respectively (6). The genus Syntrophomonas, which belonged to the family Syntrophomonadaceae, was also enriched in the AD of LCFA systems with the addition of GAC or carbon nanotubes (12, 21). It was not known whether the main LCFA-degrading bacteria enriched with HC still belonged to the above two families. Some other microbes, including Clostridium-, Thermotoga-, Coprothermobacter- and Anaerobaculum-related species have also been speculated to be possibly related to LCFA degradation (22, 23) because they were found to be enriched in the AD of LCFA degradation by 16S rRNA genes analysis. However, there was no evidence to prove that they could degrade LCFA. Reconstructing the genome of an individual strain in the mixed microbial community can explore their functional characteristics (24). The method of genome-centric metagenomics can identify metagenome-assembled genomes (MAGs) through binning the scaffolds. The functional activity of single specie displayed by genome-centric metatranscriptomics can further reveal the interaction and response mechanisms of microbial communities associated with AD (25, 26). Moreover, this method can also reveal the electron transfer mechanism by analyzing the expression of genes related to the DIET process. Therefore, the application of metagenomics and metatranscriptomics may be possible to discover more new types of bacteria with LCFA degrading ability and effectively describe the potential DIET members in the AD system.

Based on the above considerations, this study aimed to investigate the effects of HC on the methane production from LCFA, and a combination of metagenomics and metatranscriptomics analyses was applied to analyze the metabolic function changes of key strains in the community at the genetic level to discover LCFA-degrading microbes. HC might promote the degradation of LCFA through adsorption or DIET or other unclear reasons. It was known that the difficulty for the degradation of LCFA increases with the length of the carbon chain, and unsaturated LCFA was more difficult to be degraded than saturated LCFA. Therefore, this study chose oleate as the substrate, which was an unsaturated LCFA containing 18 carbons and is commonly used as a model LCFA (27).

RESULTS AND DISCUSSION

Effect of HC on the methane production from oleate.

The effect of HC on AD of oleate was investigated, and the results of methane production are shown in Fig. 1A The kinetic parameters for methane production were also calculated by the modified Gompertz model and the results are shown in Table S1.

FIG 1.

Cumulative methane production (A) and the concentration of SCFAs (B) in the first batch during the AD with and without HC. Cumulative methane production (C) and the concentration of SCFAs (D) in the second batch during the AD with and without HC.

It was found that HC effectively increased the rate of methane production (Fig. 1A). The maximum methane production rate (Rm) was increased by 56.9% with HC compared to the control experiments. In addition, HC also effectively reduced the lag phase, which was reduced from 5.7 days in the control experiments to 1.3 days (Table S1). The above results indicated that HC promoted the degradation of oleate in the AD process and eliminated the inhibitory effect of oleate. The results of the changes in SCFAs are shown in Fig. 1B Only acetate was detected in SCFAs, and the addition of HC alleviated the accumulation of acetate and promoted the methanogenesis step.

Second batch cultivation was further conducted to test whether the promotion of methane by HC was sustained and effective. Compared with the control group, the Rm of the HC group was increased by 182.7%, which was much higher than that (56.9%) in the first batch experiment. It should be noted that the Rm of the HC group was increased by 24.8% compared with the first batch cultivation of the HC group, and the time to reach the plateau phase of methane production was also shorter, which was decreased from 14 d to 12 d (Fig. 1C). At the same time, the consumption rate of acetate also became faster (Fig. 1D). The above results indicated that HC greatly enriched the microbes for oleate degradation after enrichment. However, in the second batch cultivation, the Rm of the control group was reduced by 30.7% compared with the first batch cultivation and the lag period was extended from 5.7 d to 11.5 d, which was longer than the first batch cultivation of the control group. A similar phenomenon was also observed in a previous study, which reported the methane production rate sharply decreased after cultivation in AD of LCFA (21). It indicated LCFA had strong inhibition and the activity of microbes to degrade LCFA was more inhibited after enrichment. According to the above results, it could be concluded that HC effectively eliminated the inhibitory effect of oleate on methane production and maintained the stable and efficient methane production from oleate.

HC might have absorbed oleate, which reduced inhibition and promoted methane production. Therefore, the adsorption of HC to sodium oleate was further investigated. Only 8% of oleate was adsorbed by HC (Table S2), and the decreased oleate concentration did not have obvious effects on methane production (Fig. S1 in Supplemental File 1), which showed adsorption was not the main reason for enhanced methane production from oleate by HC. To understand the reason why HC improved the methane production rate of AD of oleate, the enriched microbes, and their functions were analyzed in the following parts.

Genome reconstruction.

Metagenomic sequencing obtained 45.5 Gb of paired-end reads after quality control. After the metagenomic binning, 224 MAGs were obtained from the AD system (Fig. S2, Supplemental File 2). About 80% of the DNA reads could be mapped to these 224 MAGs. Therefore, it could be assumed that the obtained MAGs were fairly representative of the complete microbial communities in the AD system with or without HC. The 174 MAGs with completeness > 70% and contamination < 8% were selected for further study. These MAGs contained 153 bacterial and 21 archaeal MAGs, spanning 20 bacterial and 4 archaeal phyla. The bacterial MAGs were taxonomically assigned to different phyla, including Proteobacteria, Desulfobacterota, Chloroflexota, Actinobacteriota, and Firmicutes. And the archaeal MAGs were mainly assigned to Halobacteriota (13 out of 21). The principal-component analysis plot based on the data obtained from metagenomics analysis (Fig. S3A) showed that the HC group and the control group were separated, which indicated that HC significantly changed the structure of the microbial community. The specific description of changes in the microbial community was presented in Supplemental File 1.

Transcriptional responses of MAGs to the addition of HC.

Key genes involved in the LCFA degradation pathway were summarized through the Kyoto Encyclopedia of Genes and Genomes (KEGG), and the bacterial MAGs with LCFA oxidation function were selected for further research. 11 MAGs with active LCFA degradation function were screened from bacterial MAGs (Fig. 2, Table S3, Supplemental File 3). The previous studies described that the bacteria capable of degrading LCFAs belonged to the families Syntrophomonadaceae and Syntrophaceae, which belonged to the phyla Firmicutes and Proteobacteria, respectively (28). However, the selected MAGs also belonged to phyla Desulfobacterota, Spirochaetota, Chloroflexota, and Actinobacteriota besides the above two mentioned phyla. In the control group, the expression of genes in the LCFA degradation pathway of Bin190 Syntrophomonadales sp., Bin197 Syntrophomonadales sp., and Bin6 Anaerolineales sp. was upregulated compared with the HC group, which meant that the transcriptional responses of these three MAGs were more positive and responsible for oleate degradation in the control group. Bin190 Syntrophomonadales sp., Bin197 Syntrophomonadales sp., and Bin6 Anaerolineales sp. had also been proven to degrade LCFA or oftentimes appeared in the LCFA degrading systems (28, 29). With the addition of HC, Bin138 Spirochaetota sp., Bin35 Smithellaceae sp., and Bin54 Desulfomonilia sp. were identified as the dominant LCFA-degrading bacteria and displayed the high genes expression levels of this pathway, which expressed genes for more than 80% of the bioconversion steps in the LCFA-degrading pathway and significantly upregulated genes in at least three steps. This criterium was based on a previously published paper (30). Moreover, these bacteria had not been reported to degrade LCFA in the previous studies. This indicated that genome-centric metatranscriptomics was an effective method to identify more microbes with LCFA degrading functions, which enhanced the overall understanding of LCFA-degrading bacteria.

FIG 2.

Comparison of the transcriptional responses of the LCFA bioconversion-related genes in 16 bacterial MAGs in the HC group and the control group. H group and C group represented the HC group and the control group, respectively. The 21 genes determined by the KO number represented the key genes directly involved in the LCFA β-oxidation pathway, which could catalyze bioconversion in 5 steps. The heatmap was colored according to the FC and the adjusted P value of CDS for the given KO.

This study mainly focused on MAGs that responded positively after HC addition. Therefore, MAGs with increased overall RNA coverage in the HC group were further analyzed in detail. Four bacterial MAGs were significantly enriched by HC, including three LCFA-degrading bacterial MAGs: Bin138 Spirochaetota sp., Bin35 Smithellaceae sp., and Bin54 Desulfomonilia sp., and a VFA-oxidation bacterial MAG: Bin37 Syntrophobacteraceae sp. Moreover, the most positively responded archaeal MAGs in the HC group belonged to the genus Methanothrix (Fig. S4). To comprehensively understand the physiological and ecological characteristics of the four bacterial MAGs and one type of archaeal MAG, manual genome annotation and metabolic pathway reconstruction were conducted to further analyze the expression of related genes.

LCFA-degrading bacteria.

(i) The analysis of Bin138 Spirochaetota sp. In previous studies, the AD system for the degradation of LCFA-enriched bacteria belonged to the phylum Spirochaetes (the synonyms of Spirochaetota). Because the analysis of the microbial community was based on 16S rRNA gene sequencing, it was concluded that the microbes in such phylum might play an important role in the degradation of LCFA or be resistant to LCFA (31, 32). However, there was no direct and specific evidence to prove that Spirochaetes could degrade LCFA. According to the metabolic reconstruction of Bin138 Spirochaetota sp., a complete LCFA β-oxidation pathway was reconstructed (Fig. 3, Supplemental File 4).

FIG 3.

Metabolic pathways of Bin138 Spirochaetota sp. in HC group. The pathway was constructed based on the annotated genome sequences (Supplemental File 4). Black arrows indicated gene presence; purple arrows indicated gene expression.

LCFA was activated by the long-chain-fatty acyl-CoA synthetase (fadD) to fatty acyl-CoA. Then, fatty acyl-CoA was converted to enoyl-CoA, stimulated by acyl-CoA dehydrogenase (acd), and responsible for extracting two protons from fatty acyl-CoA. By introducing the water molecule into a double bond, the produced enoyl-CoA was further hydrated, forming 3L-hydroxyacyl-CoA via enoyl-CoA hydratase (fadB). Two more protons were extracted from 3L-hydroxyacyl-CoA to form β-Ketoacyl-CoA and then NADH was produced with the aid of 3L-hydroxyacyl-CoA dehydrogenase (fadN). Electrons (e⁻) were donated from NADH to H⁺, generating H₂ (28). Finally, the β-Ketoacyl-CoA, stimulated by β-Ketothiolase (fadA), was separated to acetyl-CoA and _n-2fatty acyl-CoA, and the latter was subjected to a new β-oxidation cycle. Because oleate belonged to unsaturated LCFA, after the third β-oxidation cycle, the formed _n-6fatty acyl-CoA underwent an isomerase-dependent pathway with the aid of Δ³, Δ²-enoyl-CoA isomerase (EC 5.3.3.8) to generate enoyl-CoA, and then continue the above-mentioned β-oxidation cycle (33, 34). The β-oxidation pathway worked cyclically, removing two carbon atoms from the fatty acyl-CoA in each cycle to produce acetyl-CoA and hydrogen, which were then utilized by other microbes. The acetyl-CoA was further converted to acetate with the catalysis of acetate-CoA ligase and then utilized by acetoclastic methanogens to produce methane. Compared with the control group, the expression of the enzyme genes in the LCFA β-oxidation pathway were all upregulated (8.2 to 15.8-fold) in the HC group.

According to the KEGG annotation, Bin138 Spirochaetota sp. had the butyrate oxidation pathway. The butyryl-CoA generated in the process of the LCFA β-oxidation cycle could be gradually converted into acetyl-CoA through the butyrate oxidation pathway. Additionally, Bin138 Spirochaetota sp. also encoded the gene of acetyl-CoA acetyltransferases, which cleaved acetoacetyl-CoA to generate acetyl-CoA and further transformed it into acetate. And all the genes involved in the butyrate degradation pathway in the HC group were expressed, which indicated that Bin138 Spirochaetota sp. relied on the LCFA β-oxidation pathway and used the butyrate oxidation pathway to convert butyryl-CoA into acetyl-CoA. Bin138 Spirochaetota sp. also contained genes related to central carbohydrate metabolism. It had a complete glycolysis pathway to degrade glucose (Fig. 3). Previous studies have found that the phylum Spirochaetota was enriched in the AD reactor with glucose as the substrate (35, 36). It was further confirmed that this phylum possessed this function in the present study. In addition, Bin138 Spirochaetota sp. had a complete nonoxidative branch of the pentose phosphate pathway, which produced various biosynthetic precursors, such as ribose-5-phosphate and erythrose-4-phosphate (37).

The above results showed that Bin138 Spirochaetota sp. could degrade LCFA and metabolize carbohydrates. It was a versatile microorganism dominating the microbial community, and HC promoted the positive transcriptional responses of Bin138 Spirochaetota sp. related to the degradation of LCFA.

(ii) The analysis of Bin35 Smithellaceae sp. Bin35 Smithellaceae sp. was assigned to the family Smithellaceae (phylum Desulfobacterota). Although the family Smithellaceae has not been demonstrated to be able to degrade LCFA, the genus Smithella belonging to the family Smithellaceae was found to be enriched in an anaerobic digester for degrading LCFA based on 16S rRNA gene sequencing analysis (38). Based on the metagenomic and metatranscriptomic characteristics, Bin35 Smithellaceae sp. was shown to have the ability to metabolize LCFA by the β-oxidation pathway in the present study (Fig. 4, Supplemental File 4), which provided evidence that the family Smithellaceae could degrade LCFA. HC upregulated the genes in the LCFA degradation pathway, which was about 12.8 to 15.6-fold higher than the control group. The complete metabolic pathway has been described in the chapter of Bin138 Spirochaetota sp. Moreover, this MAG might possess the ability to directly convert acetate into CO₂ (Fig. 4) because it had a relatively complete Wood-Ljungdahl pathway. Compared with the control group, the gene expression involved in the pathway of acetate conversion to CO₂ in the HC group was not upregulated, and the gene (fdha, fdhb) encoding formate dehydrogenase (NADP⁺) had no detectable expression, which indicated that HC did not affect the Wood-Ljungdahl pathway of Bin35 Smithellaceae sp.

FIG 4.

Metabolic pathways of Bin35 Smithellaceae sp. in HC group. The pathway was constructed based on the annotated genome sequences (Supplemental File 4). Black arrows indicated gene presence; purple arrows indicated gene expression; red arrows indicated gene deletion.

In addition, the genome analysis of Bin35 Smithellaceae sp. found a series of genes encoding the butyrate oxidation pathway, indicating that the MAG could utilize butyrate. The expression of genes involved in the butyrate metabolism pathway was upregulated with the addition of HC, indicating that the butyryl-CoA generated by LCFA β-oxidation pathway could continue to be oxidized through the butyrate metabolism pathway. Bin35 Smithellaceae sp. also could utilize propionate. Even though the genes encoding propionyl-CoA carboxylase and succinyl CoA synthetase were not detected, other genes involved in the propionate oxidation pathway were all present. In addition, the expression of flagellin (fliC) was significantly upregulated (11.9-fold higher than the control group), indicating that it played a vital role in fatty acid metabolism. Flagella might help microbes move forward to nutrients, contributing to expeditious synthesis or metabolism. Taken together, these observations suggested that Bin35 Smithellaceae sp. degraded LCFA both via the LCFA β-oxidation pathway and via the butyrate-oxidation pathway and could oxidize propionate.

(iii) The analysis of Bin54 Desulfomonilia sp. Bin54 Desulfomonilia sp. belonged to the class Desulfomonilia (phylum Desulfobacterota), the description of this class was the same as that of the family Desulfomonilaceae (the synonyms of Smithellaceae) (39). Previous studies have proven that two syntrophic bacteria belonging to the class Desulfomonilia possessed the butyrate oxidation pathway (30). However, no previous studies had reported that the class Desulfomonilia could degrade LCFA. Based on genome annotation, Bin54 Desulfomonilia sp. had a complete LCFA β-oxidation pathway. The expression of the genes encoding all enzymes of this pathway was upregulated in the HC group. This MAG also could metabolize butyrate, which could covert the butyryl-CoA into acetyl-CoA. The genes involved in the butyrate metabolism pathway were expressed in the HC group. The complete metabolic pathway has been described in the chapter of Bin138 Spirochaetota sp.

Previous studies have demonstrated the importance of electrically conductive pili and outer surface c-type cytochrome in the exchange of extracellular electrons involved in DIET (40). The e-pili could serve as biological nanowires in bacteria to achieve electron transfer in cell-cell interactions (41). In previous studies, empirical standards to predict whether a pilin was assembled into e-pili were established based on the e-pili conductivity mechanism (42). To prove that the gene pilA of Bin54 Desulfomonilia sp. was assembled into e-pili, we compared it with the sequence of the pilin protein of Geobacter sulfurreducens. The abundance of aromatic amino acids was 8%, closing to the 9% threshold required for high e-pili conductivity. The maximum gap between aromatic amino acids in the pilin sequence was 24 amino acids, which was far lower than the maximum aromatic-free gap (<40 amino acids). And the genes of e-pili accessory proteins (pilBCDEMOQ) were also found in this MAG. The above analysis indicated that the e-pili existed in this MAG. The genes pilABCEQ were expressed in the HC group under the stimulation of HC, but the expression was not detected in the control group. Moreover, the gene for c-type membrane cytochrome(ccsAB) also existed but was not expressed in either group. Previous studies have found that Syntrophus aciditrophicus could grow via DIET, but the genome encoded only a few putative c-type cytochromes and cytochromes were not readily apparent in cell protein (43). Not all microbes need to rely on cytochromes to efficiently transfer electrons to the outer cell surface (44). The above observations indicated that Bin54 Desulfomonilia sp. might establish a DIET network with methanogens in the HC group.

The analysis of VFA-oxidizing bacteria: Bin37 Syntrophobacteraceae sp.

Based on the genome annotation, the butyrate oxidation pathway was found in Bin37 Syntrophobacteraceae sp., and all of these genes were expressed in the HC group. Compared with the control group, the expression of this pathway was upregulated 8.19 to 12.96-fold in the HC group, indicating that Bin37 Syntrophobacteraceae sp. could further convert butyrate produced by other LCFA-degrading bacteria into acetate, which undertook the function of butyrate degradation in this system. However, the presence of butyrate was not detected, and it might be because butyrate was utilized once it was produced. In the process of butyrate degradation, acetyl-CoA before the formation of acetate could also be converted into pyruvate and acetoacetate by pyruvate ferredoxin oxidoreductase and hydroxymethylglutaryl-CoA lyase, respectively.

According to the KEGG annotation, Bin37 Syntrophobacteraceae sp. had a complete Wood-Ljungdahl pathway, and therefore it could oxidize acetate to CO₂. It had been inferred that there were some acetate-oxidizing species in the family Syntrophobacteraceae, which was analyzed by RNA-stable isotope probing technology (45). And the genes involved in the Wood-Ljungdahl pathway were expressed in the HC group. Compared with the control group, the gene expression involved in the acetate oxidation process was upregulated by 6.42 to 15.14-fold, which meant that Bin37 Syntrophobacteraceae sp. was responsible for the function of acetate oxidation in the HC group. This MAG also had a complete propionate-degradation pathway. In a previous study, Syntrophobacteraceae was the dominant microorganism in the AD system of propionate and it was speculated that this microbe was responsible for the oxidation of propionate (46). In this study, it was further verified that Syntrophobacteraceae could degrade propionate.

This result showed that HC promoted the activity of LCFA-degrading bacteria and other syntrophic bacteria to participate in the metabolism of intermediate products in this system.

The metabolic process of Methanothrix sp.

The degradation of LCFA under the syntrophy relationship of various bacteria produced a large amount of acetate, which provided an abundance of substrate for methanogens. Methanothrix sp. could not use H₂/CO₂ and formate to produce methane (47). After the addition of HC, the expression of genes in the acetoclastic methanogenesis of some Methanothrix sp. were highly upregulated, such as Bin13 Methanothrix soehngenii (2.3 to 16.3-fold higher than the control group) and Bin153 Methanothrix sp. (3.4 to 17.3-fold higher than the control group) (Fig. 5), which indicated that the HC promoted the process of converting acetate to methane by Methanothrix sp. (Supplemental File 5).

FIG 5.

Eleven archaeal MAGs involved in the process of CO₂ methanogenesis were screened. The top part of the heat map showed the transcriptional response of genes related to the hydrogenotrophic methanogenesis process. The bottom part showed the transcriptional responses of genes related to acetotrophic methanogenesis. The heatmap was colored according to the FC and the adjusted P value of CDSs for the given KO.

Generally, methanogens that utilized H₂/CO₂ required hydrogenases, including energy conversion hydrogenase (Ech), F420 reductant hydrogenase (Frh), and membrane-bound Vht (methylphenazine-reducing) hydrogenase (48). However, the genes encoding these hydrogenases were not detected in Bin13 Methanothrix soehngenii. and Bin153 Methanothrix sp., which meant that they could not convert H₂/CO₂ into methane. Nevertheless, these two MAGs possessed a complete pathway for reducing CO₂ to methane, and all enzyme-coding genes involved in this pathway were only expressed in the HC group. The results indicated that Methanothrix sp. also generated methane by reducing CO₂ with hydrogen protons and electrons, which was considered to be diagnostic for DIET (17). The above analysis revealed that HC promoted the acetoclastic methanogenesis pathway of Bin13 Methanothrix soehngenii and Bin153 Methanothrix sp. And induced their hydrogenotrophic methanogenesis pathway through DIET. Moreover, the expression of genes involved in either hydrogenotrophic methanogenesis or acetoclastic methanogenesis was upregulated in several methanogens (Fig. 5). It was inferred that the increase in methane production rate was because HC promoted the expression of the methanogenesis process of most methanogens.

The potential function of HC in DIET.

To better understand the role of HC, the effect of HC on the methanogenesis step of AD of oleate was further studied. The methanogenic activity analysis showed that when the substrate was acetate, the methane production rate of the HC group was higher than that of the control group, indicating that HC had a positive effect on acetate methanogenesis. On the other hand, when the substrate was H₂-CO₂, the methanogenic rates of the control group and HC group were not much different (Fig. S5). Moreover, the methane metabolic pathway based on metatranscriptome revealed that genes associated with hydrogenotrophic methanogenesis were upregulated with HC addition (Fig. S6, Supplemental File 5). Previous studies proved that hydrogenotrophic methanogenesis could also utilize H⁺ and e- to produce methane via DIET (7). The above results indicated that HC promoted hydrogenotrophic methanogenesis only through the process of utilizing H⁺ and e-, rather than H₂-CO₂.

The EET ability of microbes was directly related to DIET. Therefore, the effect of HC on the EET capacity of the enriched mixtures stimulated by oleate was detected by the electrochromic method (49, 50). The results of the electrochromic test displayed that the color in the test tube inoculated with the enriched mixtures with HC was darker than that in the control tube (Fig. S7). Therefore, it showed that HC promoted the extracellular electron transfer ability of microbes in the AD of oleate, which also provided evidence for the promotion of DIET by HC.

Through gene annotation of the enriched bacterial MAGs, Bin54 Desulfomonilia sp. was found to possess the e-pili genes, and these genes were only expressed in the HC group. In the process of butyrate metabolism, one pair of electrons was generated in the form of reduced electron transfer flavoprotein (ETF) during the oxidation of butyryl-CoA to crotonyl-CoA (51). This electron pair had high redox potential and could only be used by membrane complexes (such as the Fe–S oxidoreductase or the Fix complexes) (52). Bin54 Desulfomonilia sp. encoded gene clusters related to ETF, including ETF alpha-, and beta-subunits (etfA, etfB), and the expression was upregulated in the HC group. This complex could transfer electrons from reduced ETF to menaquinone with an inward proton transfer to form menaquinol and further oxidized by a membrane-bound, externally directed formate dehydrogenase or hydrogenase, accompanied by the translocation of two more protons (30). Based on the above analysis, Bin54 Desulfomonilia sp. could transfer electrons derived from butyrate metabolism to e-pili, and then might be directly used by Methanothrix sp. through DIET (Fig. 6). To more directly prove the existence of DIET, further studies of DIET in AD systems require understanding the metabolic activity and potential functions of species and need to verify their biological functions through isolation and genetic manipulation.

FIG 6.

Putative DIET model between Bin54 Desulfomonilia sp. and Methanothrix sp. as revealed by genome-centric metatranscriptomics (Supplemental File 5). Black arrows indicated gene presence; purple arrows indicated gene expression, and red arrows indicated gene deletion.

Conclusion.

The present study showed that the addition of HC increased the methane production rate from LCFA. With the integration of metagenomic and metatranscriptomic analysis, several new types of bacteria that possessed LCFA degradation function were discovered, including Bin138 Spirochaetota sp., Bin35 Smithellaceae sp., and Bin54 Desulfomonilia sp. These microbes were enriched under the stimulation of HC and are most likely related to the degradation of LCFA in the HC group. The addition of HC also promoted the change of electron transport pathway during AD of oleate, and perhaps DIET was established between Bin54 Desulfomonilia sp. and Methanothrix sp., which might also be another reason for promoting LCFA degradation and increasing methane production rate. Moreover, it was found that the LCFA was degraded to acetate through the β-oxidation pathway. The butyrate produced by LCFA-degrading bacteria via β-oxidation might also be utilized by other butyrate-oxidizing bacteria to produce acetate. Overall, this study provided new insights into the LCFA degradation with the addition of HC.

MATERIALS AND METHODS

Preparation of HC, substrate, and inoculum.

The HC was obtained by hydrothermal liquefaction. Corn straw was selected as raw material. Detailed information on HC preparation was described in Supplemental File 1. Sodium oleate was used as the substrate. The characteristics of inoculum are described in detail in Supplemental File 1.

Effects of HC on methane production.

The effect of HC on methane production was studied in batch experiments. The experiments were carried out in 118 mL serum bottles, and each bottle was filled with 60 mL fluid and 1 mL inoculum. The HC group included sodium oleate solution, inoculum, and 10 g/L HC, the control group only contained sodium oleate solution and inoculum, and the blank group only contained deionized water and inoculum. Each group was conducted in triplicates and detailed information as described in Supplemental File 1. The second batch of experiments was also conducted to see the effects of HC on methane production after the enrichment of the microbes. A similar procedure was adopted in the first batch experiment, except that the inoculum was the solid part of the bottles in the first batch experiment after centrifugation. The gas and liquid samples were collected regularly for analysis, and the analytical methods for methane content and VFA concentrations were described in Supplemental File 1. The modified Gompertz model was also used to simulate methane production (Supplemental File 1). The extracellular electron transfer (EET) capability of the enriched microbes was evaluated by the color change experiment of WO₃ nanoclusters (Supplemental File 1). To better understand the promoting effect of HC adsorption on oleate methane production, the ability of HC to adsorb sodium oleate was also measured (Supplemental File 1).

Genome-centric metagenomic analyses.

For metagenomic analysis, samples (20 mL) were collected from three parallels in the control and HC groups during the active methanogenesis phase of the second batch culture (day 7 for the HC group and day 22 for the control group). The collected samples were centrifuged at 10,000 rpm for 10 min at 4°C to discard the supernatant and collected about 1 g of the cell pellet. Each sample was divided into two aliquots (approximately 1 g weight each), of which one was subjected to DNA extraction for metagenomic analysis and the other one was subjected to RNA extraction for metatranscriptomics analysis. PowerMax Soil DNA isolation kit (MoBio Laboratories, USA) was used to extract the total genomic DNA of all the samples. The Illumina Nexttera DNA XT kit (Illumina, San Diego, CA, USA) was used to prepare a metagenomic library (2 × 150 bp). The Trimmomatic software (v0.38) and the MEGAHIT (v1.0) were used to filter, trim, and assemble the raw metagenomic sequences, respectively (53, 54). Metagenomic binning was applied to the coassembly according to to CONCOCT, MetaBAT2 and MaxBin2, and the generated bins were consolidated into a single bin set using the MetaWRAP pipeline to produce the final MAGs. The CheckM (v1.0.18) (55) was used to estimate the completeness and contamination of each MAG and Prokka (v1.13.7) (56) was used to annotate the MAGs with the bacterial or archaeal mode. The KEGG was applied to annotate the coding sequences (CDS) data set of MAG, and the KEGG orthologous group ids (KO) were used for metabolic pathway analysis (57). A detailed description is provided in Supplemental File 1.

Genome-centric metatranscriptomics analysis.

As described in metagenomics analysis, samples from the active metabolism phase of the second batch culture were also used for metatranscriptomics analysis. The RNeasy PowerMicrobiome kit (Qiagen, Germany) was used to extract total RNA, and the DNase Max kit (Qiagen, Germany) was used for DNase treatment to remove DNA contaminants from the extracted RNA. RNA integrity and quality were checked by Agilent Bioanalyzer (Agilent, USA) and the RNA integrity number was higher than 6.

For these 6 samples, the Ribo-Zero rRNA Removal kit (Illumina, San Diego, CA, USA) was used to remove rRNA from the total RNA. According to the manufacturer's instructions, the TruSeq RNA Sample Prep kit (Illumina, San Diego, CA, USA) was applied to prepare cRNA sequencing libraries. HiSeq3000/4000PE Cluster kit (Illumina Inc.) was used to sequence single-end (50 bp) of the prepared libraries on a HiSeq3000/4000 SBS. The Trimmomatic (v0.38) was used to trim the raw RNA reads to remove adaptors and low-quality reads with a Phred value <20 and read length <50 bp. Then through SortMeRNA (v4.2.0), the trimmed RNA reads were aligned with the reference database to filter the rRNA reads, and the unmapped reads were treated as mRNA sequences (58). The Bowtie 2 was used to map mRNA reads to the CDS reference data set made by the constructed MAG, and then Samtools (v1.0) was used for ordering and indexing (2, 59). CheckM was used to determine the transcriptomic data and Htseqcount (v0.6.1) was used to calculate the read count for each CDS, normalized by the sequencing depth and CDS length, which was expressed as reads per Kb-CDS per million mapped reads (RPKM) values (60).

Genome-centric metatranscriptomics analysis determined the highly expressed CDS in each genome, which indicated enzyme activity. The coverages of MAGs in metagenome and metatranscriptome were obtained by calculating the number of DNA or mRNA reads aligned with the contigs or CDS of the homologous MAG. Furthermore, each MAG was statistically analyzed using edgeR software (61). The P value and fold change (FC) were calculated for the expression of individual CDS. Benjamini and Hochberg's method was used to correct the P value and obtained the adjusted P value to eliminate the possibility of a false-positive. The classification of genes expression was as follows: upregulated (HC group): FC ≥ 2 and adjust P < 0.05; not regulated (HC group): 0.5 < FC < 2 or adjust P > 0.05; downregulated (HC group): FC ≤ 0.5 and adjust P < 0.05. To determine the influence of the genome abundance variation on the gene expression, the same reads alignment, and filtering procedure was applied utilizing the DNA reads obtained from three parallel samples in each group (62). The change in the DNA coverage between the two groups (with or without HC) was determined by changes in the species abundance in the microbial community. Comparing the results acquired from the edgeR software determined for DNA reads and RNA reads, the influence of genomic abundance variation was removed from the changes in gene expression (25).

Data availability.

The raw sequencing data had been submitted to NCBI Sequence Read Archive under the project ID PRJNA850144.